Literature Review

According to a study conducted by (Raji et al., 2024) illustrates that in today’s highly competitive e-commerce landscape, understanding customer behavior is crucial for business success. In support of this, (Gold et al., 2024) highlights that data analytics has emerged as a powerful tool that enables companies to identify distinct customer segments, allowing for personalized marketing strategies that resonate with individual preferences. As per the study conducted by (Sheed Iseal & Michael, 2025) highlighted that, Amazon company implemented advanced data analytics to evaluate purchase history, browsing patterns, and customer reviews. Also, supporting this (Rohit Guddadakeri Shivanand et al., 2023) research illustrates that advanced data analytics help Amazon to recommend products to users according to their preferences, which significantly enhances the number of sales. According to research conducted by (Horton, 2025), the effective customer segmentation with the help of data analytics not only enhances user experience it also drives customer loyalty which is very beneficial for the company. (Bough et al. ,2023), illustrates that companies that employ customer analytics to gain insight and organize their customers into demographic, purchasing, and engagement segments experience an increase in revenue of 10-15% and a 20% increase in customer retention. Making informed decisions through segmentation and customer analytics, businesses can design targeted marketing programs that enhance the relevance and effectiveness of the marketing communications (Hoenig, 2025).  For example, Nike uses data analytics to promote specific products to different customer groups, increasing the likelihood of purchase and brand loyalty(Patov, 2024). Furthermore, (Le, 2024) demonstrates that data-driven segmentation allows e-commerce businesses to optimize their advertising spend. Instead of broad, untargeted campaigns, companies can focus on high-value segments, maximizing return on investment (AI, 2024). As e-commerce continues to grow projected to reach $6.3 trillion globally by 2024 leveraging data analytics becomes essential for gaining a competitive edge(Pradhan, 2024). 

Research Gap

Despite the extensive use of data analytics in customer segmentation, there remains a gap in understanding how emerging technologies like artificial intelligence and machine learning can further enhance segmentation accuracy and predictive capabilities. Additionally, many studies focus on large corporations, leaving a gap in insights for small and medium-sized enterprises (SMEs), which may face resource constraints. Addressing these gaps could lead to more inclusive and advanced segmentation strategies that benefit a broader spectrum of e-commerce businesses.

Research Methodology

Approaches and Methods:

This research adopts a qualitative approach, utilizing case studies and expert interviews to gain in-depth insights into the application of data analytics in customer segmentation within the e-commerce sector. The focus is on understanding the processes, challenges, and benefits associated with data-driven segmentation strategies from industry professionals’ perspectives.  

Data Collection, Sampling, and Analysis

Data will be collected primarily through purposive sampling, targeting marketing professionals, data analysts, and industry experts with experience for e-commerce and data analytics. Secondary data extracts will be taken from industry reports, academic journals, company websites, as well as market research databases. Data from interviews and case study will be qualitatively researched and analyzed through a thematic analysis in order to extract themes, issues and insight concerning the utilization of data analytics in customer segmentation.

Ethical Issues  

The research will ensure confidentiality and anonymity of interviewees, obtaining informed consent prior to data collection. Any primary data gathered will be used solely for academic purposes, and sources will be appropriately cited to avoid plagiarism. The study will also adhere to ethical standards related to data privacy and intellectual property rights.

Expected Sources of Information  

The primary sources consist of industry reports produced by analysts such as Statista, McKinsey, and eMarketer as well as academic articles from databases like IEEE Xplore, and Google Scholar. Secondary sources also include company websites and white papers that add specific context-related information. These sources will provide sufficient data to analyze data analytics in the context of customer segmentation in e-commerce.

Potential outcomes

This research is expected to provide insights on how data analytics can improve customer segmentation through the e-commerce sector. We argue to demonstrate that machine learning and artificial intelligence can not only help improve those aspects of segmentation accuracy, value, and predictive power. The results may provide the current best practices of segmentation techniques employed by companies that are successful in e-commerce with their customer segmentation, effective personalization, enhancing customer engagement, and the touchpoint strategies employed with their general customer experience. In addition, the results may fill the gaps in understanding how new emerging technologies can be integrated with existing traditional or data-informed segmentation reference models used by small and medium-sized enterprises. More broadly, the practical implications provided from the study may help companies employ data analytics more effectively to capitalize on the benefits of customer retention, sales revenue, and resource allocation efficiency. The research is also designed to contribute to academic content by validating and confirming or rejecting existing theories on data-informed customer segmentation models in order to fill the gaps on data-informed segmentation processes that are used by entrepreneurs beyond the realm of consumerism. The broader results of this research can lead practitioners and academics towards progressive, more innovative, and inclusive segmentation practices.

References

BYYD. (2025, May 6). BYYD. https://www.byyd.me/en/blog/2025/05/how-e-commerce-will-evolve-in-2025-key-predictions/

Bamidele Micheal Omowole, Amarachi Queen Olufemi-Phillips, Onyeka Chrisanctus Ofodile, Nsisong Louis Eyo-Udo, & Somto Emmanuel Ewim. (2024). Big data for SMEs: A review of utilization strategies for market analysis and customer insight. International Journal of Scholarly Research in Multidisciplinary Studies, 5(2), 001-018. https://doi.org/10.56781/ijsrms.2024.5.2.0044

Charlotte, L. J. (2020, July 17). Data-Driven Customer Segmentation and Personalization in E-Commerce. Researchgate. https://www.researchgate.net/publication/390873255_Data-Driven_Customer_Segmentation_and_Personalization_in_E-Commerce

Gold, N., Egieya, N. Z. E., Ikwue, N. U., Udeh, A., Adaga, M., DaraOjimba, D., & Osato, N. (2024). Leveraging Big Data for Personalized Marketing campaigns: a Review. International Journal of Management & Entrepreneurship Research, 6(1), 216–242. https://doi.org/10.51594/ijmer.v6i1.778

Lindecrantz, E., Gi, M. T. P., & Zerbi, S. (2020, April 28). Personalized experience for customers: Driving differentiation in retail | McKinsey. Mckinsey & Company. https://www.mckinsey.com/industries/retail/our-insights/personalizing-the-customer-experience-driving-differentiation-in-retail

Majeed, M., Chaudhary, A., & Chadha, R. (2024, December 6). Digital Transformation in the Customer Experience. Researchgate. https://doi.org/10.1201/9781003560449

Ogun, V., Willian, L., & Ogunrinde, V. (2025, January 5). The Post-COVID-19 Shopping Experience: Exploring the Role of Emerging Retail Technologies. Researchgate. https://www.researchgate.net/publication/387741547_The_Post-COVID-19_Shopping_Experience_Exploring_the_Role_of_Emerging_Retail_Technologies

Prosper, J. (2024, June 1). Data Analytics and Machine Learning for Omni- Channel Optimization. Researchgate. https://www.researchgate.net/publication/389651028_Data_Analytics_and_Machine_Learning_for_Omni-_Channel_Optimization

Raji, M. A., Olodo, H. B., Oke, T. T., Addy, W. A., Ofodile, O. C., & Oyewole, A. T. (2024). E-commerce and Consumer behavior: a Review of AI-powered Personalization and Market Trends. GSC Advanced Research and Reviews, 18(3), 066–077. https://doi.org/10.30574/gscarr.2024.18.3.0090

Rohit Guddadakeri Shivanand, Zhang, Z., Duan, J., & Liu, Y. (2023, December 11). Using Amazon as a case, a mixed-method study to explore the impact of Personalised Recommendation systems on User Experience and Decision-making. ResearchGate. https://doi.org/10.13140/RG.2.2.13685.96486

Sheed Iseal, & Michael, H. (2025, February 7). Customer Behavior Analysis and Purchase Prediction in E-Commerce. Researchgate. https://www.researchgate.net/publication/388779959_Customer_Behavior_Analysis_and_Purchase_Prediction_in_E-Commerce

Statista. (2025, June 23). Areas with most use for data-driven marketing 2022. Statista. https://www.statista.com/statistics/264143/marketing-areas-influenced-by-data-and-audience-information-worldwide/

Swinscoe, A. (2023). 15 Customer Experience Predictions For 2024. Forbes. https://www.forbes.com/sites/adrianswinscoe/2023/12/18/15-customer-experience-predictions-for-2024/

Tavor, T., Gonen, L. D., & Spiegel, U. (2023). Customer Segmentation as a Revenue Generator for Profit Purposes. Mathematics, 11(21), 4425. MDPI. https://doi.org/10.3390/math11214425

Theodorakopoulos, L., & Theodoropoulou, A. (2024). Leveraging Big Data Analytics for Understanding Consumer Behavior in Digital Marketing: a Systematic Review. Human Behavior and Emerging Technologies, 2024(1). https://doi.org/10.1155/2024/3641502

Introduction

The digital world sees people and organizations produce an overwhelming amount of textual and visual content on a daily basis, ranging from social media updates and learning materials to customer reviews and research repositories (Dwivedi et al., 2021). Although this data explosion creates possibilities for more profound understanding and engagement, it also poses the challenges of interpreting, structuring, and making these diverse content forms meaningful. Most current tools support either text or image analysis but not in combination, resulting in disconnected insights and lost potential for richer, context-aware understanding.  

With industries shifting towards personalization and data-driven decision-making, there is an increasing demand for platforms that can efficiently analyze and synthesize multimodal data in order to enable users to discover latent patterns and derive meaningful insights (Rashid and Kausik, 2024). Moreover, conventional analysis techniques are usually found lacking in being able to provide real-time responsiveness and dynamism, which are critical in dynamic digital settings (Theodoropoulou and Stamatiou, 2024).  

In this context, technological advancement in artificial intelligence and machine learning have made it possible to create new interactive systems that can simultaneously examine images and text (Case Western Reserve University, 2024). This project stands in this changing context, seeking to create an interactive content generation platform that integrates image and text analysis to drive content comprehension, generation, and user interaction.

Research Question

How might multimodal machine learning methods be tailored to create an interactive platform that efficiently integrates image and text processing for dynamic content creation?

AIM

To develop an interactive multimodal content generation platform that seamlessly integrates image and text analysis to enhance real-time content creation, understanding, and user engagement.

Research Objectives

In order to compare and assess the efficacy of current state-of-the-art multimodal machine learning models for combined image and text analysis.

• To create and build an interactive platform that facilitates dynamic content creation by integrating visual and text data.
• To architect the platform for real-time processing, to support scalability and responsiveness for various user applications.
• To evaluate the quality, coherence, and relevance of created content using standard test metrics and user ratings.
• To examine the real-world practical problems, constraints, and moral concerns related to installing multimodal content generation systems in actual settings.
• To offer practical suggestions to increase user participation and content comprehension through the implementation of sophisticated multimodal AI methods.

Brief Literature Review

According to (Dwivedi et al., 2021), the rise of digital platforms has transformed everyday interactions into a flood of images and words, each carrying valuable signals waiting to be unlocked. (Dwivedi et al., 2021) illustrate that traditional content analysis tools often treat images and text as separate silos, which limits the understanding of the deeper stories they tell when combined. (Shevgan, 2025) demonstrates that as industries pursue hyper-personalization, there is a growing realization that the true power of content analysis lies in understanding these modalities collectively, rather than in isolation.  

New developments in artificial intelligence allow for the simultaneous interpretation of images and text, adding a richer context to content analysis (Case Western Reserve University, 2024). (Clarifai, 2025) highlight that platforms like Clarifai and Google Vision AI leverage deep learning for object recognition, sentiment analysis, and contextual interpretation. (Pranjić, 2020) demonstrate that combining image and text analysis can boost user engagement by as much as 40% and enhance analytical accuracy by 25-35%, illustrating the practical benefits of this integrated approach.  

Still, challenges remain in maximizing these systems for real-time response and user-friendly interfaces, while also addressing issues related to data privacy and bias (Theodoropoulou and Stamatiou, 2024). (Kumar and Subramani R, 2024) highlight that researchers emphasize the importance of user-centric design, which balances powerful AI capabilities with seamless, frictionless interaction to unlock the full potential of multimodal analysis. (Gligorea et al., 2023) demonstrate that engaging with these tools provides rich insights for learners and developers, influencing the skills needed to develop innovative AI-based systems.  

This project is situated at the intersection of multimodal AI and human-centered design, as it aims to extend the frontiers of content generation and analysis by creating an interactive platform that combines visual and text-based intelligence to produce informative, customized digital experiences.

Research Methodology

This research will take a quantitative, experimental methodology to design, deploy, and test an interactive platform that couples image analysis and text analysis for generating dynamic content.

Data Collection and Preprocessing:

Paired images and text annotations datasets (e.g., COCO, Flickr8k) will be obtained to provide multimodal data diversity. Normalization, resizing, and augmentation will be applied to images and preprocessed via tokenization, stopword removal, and embedding for textual data in order to facilitate proper model training.

The platform will be constructed based on Python, TensorFlow, and PyTorch, utilizing CNNs and Vision Transformers for vision and transformer-based NLP models such as BERT and GPT for text to perform image analysis and text analysis. Multimodal data fusion methods will be used to merge visual and text features to support coherent content creation.

Content Generation and System Integration

Higher generative models (image GANs, text transformer-based language models) will be incorporated into the platform to allow users to create new content, captions, or textual abstracts dynamically.

Training and Tuning:

Models will be trained with gathered datasets with hyperparameter tuning through grid search or Bayesian optimization to attain the best accuracy, coherence, and responsiveness.

Evaluation:

Performance will be measured in terms of metrics like BLEU, ROUGE, and CIDEr for text outputs and accuracy and F1-score for image recognition. Feedback from the users will be gathered to gauge usability, relevance, and level of engagement.

Ethical Considerations:

Data privacy will be ensured through anonymization and secure storage, with transparent user consent. Techniques to reduce bias, such as balanced datasets and fairness filters, will be implemented. Regular audits and user feedback will help address ethical issues, ensuring the platform promotes responsible, equitable, and transparent AI content generation aligned with best practices.

Outcome Alignment:

This approach will facilitate the structured creation and experimentation with a scalable, user-friendly interactive content generation platform, driving the research goals of improving multimodal AI integration for real-world content analysis and generation.

Potential Outcomes

This project expects that the created platform will show how multimodal machine learning models, integrated together, can greatly improve digital content analysis and generation, fusing visual and textual information to generate consistent, pertinent, and engaging results. Leveraging CNNs, Vision Transformers, and NLP models such as BERT and GPT, the system should outperform existing single-modality tools in accuracy, coherence, and analysis depth.

By user testing and performance benchmarking, the project will measure user engagement and content comprehension improvements quantitatively, confirming research outcomes like a possible 25–35% analytical accuracy boost when image and text analysis are integrated (Pranjić, 2020).

In addition, the project will make a contribution to the literature by providing comparative analysis of the performance of various multimodal models in actual deployments, informing future research and practical application in content analysis. It will also identify top challenges of real-time processing, scalability, and bias reduction, with concrete advice for practitioners interested in deploying interactive multimodal AI systems in digital education, social media, and marketing settings.

Research Plan & Timeline

Reference list

Case Western Reserve University (2024). Advancements in Artificial Intelligence and Machine Learning. [online] CWRU Online Engineering. Available at: https://online-engineering.case.edu/blog/advancements-in-artificial-intelligence-and-machine-learning.

Cen, Z. and Zhao, Y. (2024). Enhancing User Engagement through Adaptive Interfaces: A Study on Real-time Personalization in Web Applications. Enhancing User Engagement through Adaptive Interfaces: A Study on Real-time Personalization in Web Applications, [online] 1(6), pp.1–7. doi:https://doi.org/10.70393/6a6574626d.323332.

Clarifai (2025). NLP API | Pre-trained NLP Models for Text & Image Data. [online] Clarifai.com. Available at: https://www.clarifai.com/products/natural-language-processing [Accessed 4 Jul. 2025].

Dwivedi, Y.K., Ismagilova, E., Hughes, D.L. and Carlson, J. (2021). Setting the future of digital and social media marketing research: Perspectives and research propositions. International Journal of Information Management, [online] 59(1), pp.1–37. doi:https://doi.org/10.1016/j.ijinfomgt.2020.102168.

Gligorea, I., Cioca, M., Oancea, R., Gorski, A.-T., Gorski, H. and Tudorache, P. (2023). Adaptive Learning Using Artificial Intelligence in e-Learning: A Literature Review. Education Sciences, 13(12), pp.1216–1216.

Kumar, N. and Subramani R (2024). A Descriptive Study on Emerging AI Tools in Digital Media Content Creation. International Journal of Research Publication and Reviews, [online] 5(11), pp.1447–1452. Available at: https://www.researchgate.net/publication/385810667_A_Descriptive_Study_on_Emerging_AI_Tools_in_Digital_Media_Content_Creation.

Pranjić, G. (2020). Proceedings of FEB Zagreb International Odyssey Conference on Economics and Business, 2020(1). doi:https://doi.org/10.22598/odyssey/2020.2.

Rashid, A.B. and Kausik, A.K. (2024). AI Revolutionizing Industries Worldwide: a Comprehensive Overview of Its Diverse Applications. Hybrid Advances, [online] 7(100277), pp.100277–100277. doi:https://doi.org/10.1016/j.hybadv.2024.100277.

Shevgan, M. (2025). Content Management System Market Size & Forecast, 2025-2032. [online] Coherent Market Insights. Available at: https://www.coherentmarketinsights.com/industry-reports/content-management-system-market [Accessed 4 Jul. 2025].

Theodorakopoulos, L., Theodoropoulou, A. and Stamatiou, Y. (2024). A State-of-the-Art Review in Big Data Management Engineering: Real-Life Case Studies, Challenges, and Future Research Directions. Eng, [online] 5(3), pp.1266–1297. Available at: https://www.mdpi.com/2673-4117/5/3/68.

Background Information

Project performance evaluation is the process of assessing how well a project is progressing and whether it is achieving its intended goals (Zwikael & Meredith, 2019) . In this context, (Kashiwagi & Byfield, 2007) discussed that, in a construction project, this may simply involve checking whether the work was completed on time, on budget, and to the required quality standard. Similarly, (Meredith & Zwikael, 2019) mentioned that, in a software development project, evaluation may include tracking whether the project is on time, on budget, and delivers the features stated in the project plan. According to (Shah et al., 2023), organizations with effective project evaluation processes are 28 % more likely to complete their projects on time and within budget. (Michael Osinakachukwu Ezeh et al., 2024) explains there are numerous factors impacting the success of projects, including efficiency of team members, stakeholder management, allocation of resources (money, funds and materials), and risk management. In addition, (Bucăţa & Rizescu, 2017) indicate the proper flow of communication is essential for the organizations or businesses to keep everyone aligned. Historically, (Varma et al., 2023) explains, organizations relied on manual methods to assess these factors, such as reading reports, having meetings, and human judgment. However, as the number of projects and their complexity increase, these manual methods can become slow, prone to errors, and less reliable (Marle & Vidal, 2015). (Torres et al., 2021) also argued that evaluations of projects manually can take up to 50% longer and are more prone to mistakes. This has led to the adoption of new tools and technologies that can automatically monitor and assess project performance in real time (Vergara et al., 2025).  

Recently, (Gligorea et al., 2023) demonstrated that data analytics has become a vital component in project performance evaluation. Advanced analytics techniques enable organizations to explore large datasets, identify patterns, and generate insights that support more informed decision-making. Industry reports published by (Salimimoghadam et al., 2025) indicate that leveraging data analytics can significantly reduce evaluation time up to 50% while improving accuracy and enabling proactive risk management. These capabilities illustrate how data-driven approaches can enhance the speed and reliability of project assessments. However, despite these promising benefits, organizations face notable challenges. (Samuel Omokhafe Yusuf et al., 2024) highlight concerns about high costs of implementing sophisticated analytics systems, and a shortage of skilled personnel to operate and interpret data effectively.

In this concern, (Hayat,2025) suggests that, depending on various components, including size and the complexity of the project, piloting an advanced data analytics solution can range from a few thousand to several million dollars for an organization. These uncertainties highlight the need to fully comprehend the potential benefits and barriers involved in the adoption of data analytics for the evaluation of projects before committing resources. Also, (Sánchez et al., 2025) point out that the current dynamic business world requires organizations to have a reliable way to measure project performance. This underscores the pressing need to investigate how project performance management can be successfully fused with data analytics to drive competitive advantage and successful project outcomes (Nyathani, 2023).

Research Aim

To investigate and explore organizational challenges and opportunities in utilizing data analytics for evaluating project performance. Based on this, viable strategies and solutions will be recommended for improving project management practices through the application of technology.

Research Questions:

RQ1. What data analytics techniques (such as descriptive, diagnostic, predictive, and prescriptive analytics) are most effective in transforming the accuracy, timeliness, and comprehensiveness of project performance evaluations compared to traditional manual methods?

RQ2. How can advanced data analytics methods, including machine learning and deep learning techniques, be applied to improve project performance evaluation such as predicting delays, identifying risks, and optimizing resource allocation?

Research Objectives: 

• To evaluate the capabilities and limitations of various data analytics techniques including descriptive, diagnostic, predictive, and prescriptive analytics in automating and enhancing the accuracy, timeliness, and predictive power of project performance assessments.
• To determine how advanced techniques in data analytics (ML, DL, etc.) can enhance the existing project evaluation process such as predicting delays, identifying risks, and allocational optimization.
• To identify organisational, technical, ethical and resource barriers to establishing analytics systems to manage project performance and suggest approaches to mitigating these barriers.
• To implement strategic guidelines for organisations who wish to use and leverage data analytics tools in project management, aiming the maximum benefit from it and tackling the present challenges.

Literature Review

Project performance is a critical measure of an organization’s success, reflecting how well resources are utilized to meet goals within specified timelines and budgets (Ahmed, 2023). A study conducted by (Ofer Zwikael & Jack R. Meredith, 2019) demonstrated that effective evaluation of project performance is essential for identifying areas of improvement, ensuring accountability, and guiding strategic decision-making. In this regard, (Muftah, 2022) illustrated that traditionally, it relied on manual methods such as progress reports, financial audits, and qualitative assessments/organizations to gauge project outcomes. (Khahro et al., 2023) emphasized that while these methods provided valuable insights, they often suffered from delays, subjectivity, and limited scope, making timely decision-making challenging. (Molina & Gregson, 2017) demonstrated that in recent years, there has been a growing recognition of the need for more accurate and real-time evaluation methods. This has led to the integration of data analytics techniques into project management processes (Mortaji & Shateri, 2024). (Wolniak & Grebski, 2023) exemplified that descriptive analytics helps organizations understand historical data, while diagnostic analytics uncovers reasons behind certain outcomes. (Adesina et al., 2024) highlighted that predictive analytics forecasts potential risks and delays, enabling proactive responses, and prescriptive analytics offers recommendations for optimal decision-making. (Odejide & Edunjobi, 2024) elaborated that the advent of advanced technologies like machine learning and deep learning has further enhanced these capabilities by allowing systems to identify complex patterns, predict future project performance, and automate routine evaluation. According to a recent report conducted by (Ogunbukola, 2024) highlighted that organizations adopting predictive analytics in project management experienced a 20% reduction in project delays and a 15% increase in resource efficiency, demonstrating their potential to transform project evaluation. Despite these advancements, organizations face several challenges when implementing data analytics solutions (Medida & Kumar, 2024). Moreover, (Omokhafe et al., 2024) demonstrated that data quality and integration issues often hinder analytics accuracy, while the high costs of technology adoption and skilled personnel can be prohibitive. As per (Pina et al., 2024), ethical concerns surrounding data privacy and security also pose significant barriers. Moreover, (Damawan & Azizah, 2020) illustrated that resistance to change within organizations and lack of strategic alignment can impede the successful deployment of these tools. Overcoming these challenges requires a clear strategy, investment in training, and a focus on ethical data use to fully harness the benefits of data analytics in project performance management (Abdul-Azeez et al., 2024).

Research Gap

Although data analytics holds significant promise for enhancing the evaluation of project performance, several research gaps still exist (Vicci, 2024). Most studies (Raftopoulos & Juho Hamari, 2024) recognize the technical capabilities of data analytics tools and organizational readiness; however, they often overlook how organizations operate, the specific challenges they encounter, or the strategies necessary for effective implementation. Empirical research exploring how organizations, particularly small and medium-sized enterprises (SMEs) develop strategies to address the costs associated with adopting advanced analytics and overcome barriers such as data security concerns, ethical considerations, and skills shortages remains limited (Zavodna et al., 2024). Furthermore, there is a lack of empirical studies documenting best practices for integrating data analytics into existing project management frameworks or understanding how organizational culture influences the adoption process. These gaps hinder the development of practical, context-specific strategies for deploying data analytics in diverse organizational settings. Closing these research gaps is essential to facilitate a smoother integration of data analytics tools, ultimately enabling organizations to leverage real-time insights, improve decision-making, and enhance overall project performance effectively.

Research Methodology

Approach and Methods: 

The methodology of this study is qualitative, emphasizing an examination of existing research and theoretical literature concerning the factors influencing project evaluation and the integration of data analytics within project management. Since the review does not involve quantitative data analysis techniques or software such as NVivo, it primarily relies on a comprehensive review of industry reports, case studies, and scholarly articles to gather insights.

Data Collection: 

Data sources are drawn from an extensive review of peer-reviewed journals, books, conference papers, industry white papers, and reputable online platforms. The selection criteria prioritize recent literature (preferably within the last five years), relevance to data analytics in project evaluation, and the credibility of the sources to ensure the validity and reliability of the information.

Sampling: 

A purposive sampling approach is employed to select sources that specifically address factors impacting project evaluation and the application of data analytics. The focus is on high-impact journals, industry reports, and case studies from organizations that have successfully implemented data analytics tools in their project management practices.

Data Analysis: 

Sources of information include scholarly journals such as the International Journal of Project Management, industry white papers from organizations like PMI and McKinsey, relevant books, and reputable online databases like Google Scholar and ResearchGate. The analysis involves synthesizing insights from these sources to identify common challenges, best practices, and strategic considerations related to the adoption and integration of data analytics in project evaluation.

Tools/Resources/Software Table

Potential Outcomes

Based on the review of existing literature and the qualitative analysis conducted, several key outcomes are anticipated. Initially, the research is expected to reaffirm the continued importance of traditional project evaluation factors, such as resource allocation, stakeholder mapping, risk assessment, and effective communication. The study will likely demonstrate that these evaluation components remain critical to project success, even as the methods and tools used for evaluation evolve with technological advancements.Secondly, the research is expected to show that data analytics significantly enhances the quality and effectiveness of project evaluation. For example, data analytics tools utilizing big data and real-time data processing can provide timely insights, predict project outcomes, and support pre-decision making, thereby improving overall project success (Nabeel, 2024). For instance, organizations employing data analytics have reported improvements in project outcomes and reductions in evaluation time by up to 50% (Celestin & N. Vanitha, 2017).   In summary, it is probable that the findings will illustrate how data analytics can uncover hidden patterns and generate insights that are not easily attainable through traditional manual evaluation methods. This analysis will contribute to the body of knowledge by emphasizing the importance for organizations to adopt data-driven approaches within their project management practices. Ultimately, the research will identify gaps in the application of data analytics such as organizational resistance, lack of awareness of available technologies, or insufficient skills highlighting areas for future research and practical implementation opportunities.

Reference list

Raftopoulos, M., & Juho Hamari. (2024). Organizational Challenges in Adoption and Implementation of Artificial Intelligence. Hawaii International Conference on System Sciences (HICSS). https://www.researchgate.net/publication/379831873_Organizational_Challenges_in_Adoption_and_Implementation_of_Artificial_Intelligence

Vicci, Dr. H. (2024). The Power of Artificial Intelligence in Project Management. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.4834838

Zavodna, L. S., Überwimmer, M., & Frankus, E. (2024). Barriers to the implementation of artificial intelligence in small and medium sized enterprises: Pilot study. Journal of Economics and Management, 46(1), 331–352. https://doi.org/10.22367/jem.2024.46.13

Budhwar, P. . . .Chowdhury, S. . . .Wood, G. . . .Aguinis, H. . . .Bamber, G. J.. . .Beltran, J. R.. . .Boselie, P. . . .Lee Cooke, F. . . .Decker, S. . . .DeNisi, A. . . .Dey, P. K.. . .Guest, D. . . .Knoblich, A. J.. . .Malik, A. . . .Paauwe, J. . . .Papagiannidis, S. . . .Patel, C. . . .Pereira, V. . . .Ren, S. . . .Rogelberg, S. Varma, A. (2023). Human resource management in the age of generative artificial intelligence: Perspectives and research directions on ChatGPT Human Resource Management Journal, 33(3), 606-659. https://doi.org/10.1111/1748-8583.12524

Marle, F., & Vidal, L.-A. (2015). Limits of Traditional Project Management Approaches When Facing Complexity Managing Complex, High Risk Projects, 53-74. https://doi.org/10.1007/978-1-4471-6787-7_2

Torres, Y., Nadeau, S., & Landau, K. (2021). Classification and Quantification of Human Error in Manufacturing: A Case Study in Complex Manual Assembly Applied Sciences, 11(2), 749. https://doi.org/10.3390/app11020749

Vergara, D., del Bosque, A., Lampropoulos, G., & Fernández-Arias, P. (2025). Trends and Applications of Artificial Intelligence in Project Management Electronics, 14(4), 800. https://doi.org/10.3390/electronics14040800

Hirani, R., Noruzi, K., Khuram, H., Hussaini, A. S., Aifuwa, E. I., Ely, K. E., Lewis, J. M., Gabr, A. E., Smiley, A., Tiwari, R. K., & Etienne, M. (2024). Artificial Intelligence and Healthcare: A Journey through History, Present Innovations, and Future Possibilities Life, 14(5), 557. https://doi.org/10.3390/life14050557

Gligorea, I., Cioca, M., Oancea, R., Gorski, A.-T., Gorski, H., & Tudorache, P. (2023). Adaptive Learning Using Artificial Intelligence in e-Learning: A Literature Review Education Sciences, 13(12), 1216. https://doi.org/10.3390/educsci13121216

Salimimoghadam, S., Ghanbaripour, A. N., Tumpa, R. J., Kamel Rahimi, A., Golmoradi, M., Rashidian, S., & Skitmore, M. (2025). The Rise of Artificial Intelligence in Project Management: A Systematic Literature Review of Current Opportunities, Enablers, and Barriers Buildings, 15(7), 1130. https://doi.org/10.3390/buildings15071130

Țîrcovnicu, G., Hațegan, D., (2023). INTEGRATION OF ARTIFICIAL INTELLIGENCE IN THE RISK MANAGEMENT PROCESS: AN ANALYSIS OF OPPORTUNITIES AND CHALLENGES Journal of Financial Studies, 8(15), 198-214. https://doi.org/10.55654/jfs.2023.8.15.13

Samuel Omokhafe Yusuf, Remilekun Lilian Durodola, Godbless Ocran, Justina Eweala Abubakar, Amarachi Zita Echere, & Adedamola Hadassah Paul-Adeleye, (2024). Challenges and opportunities in AI and digital transformation for SMEs: A cross-continental perspective World Journal of Advanced Research and Reviews, 23(3), 668-678. https://doi.org/10.30574/wjarr.2024.23.3.2511

Zavodna, L. S., Überwimmer, M., Frankus, E., (2024). Barriers to the implementation of artificial intelligence in small and medium sized enterprises: Pilot study Journal of Economics and Management, 46331-352. https://doi.org/10.22367/jem.2024.46.13

Überwimmer, M. S., Zavodna, L., Frankus, E., (2024). Barriers to the implementation of artificial intelligence in small and medium sized enterprises: Pilot study Journal of Economics and Management, 46331-352. https://doi.org/10.22367/jem.2024.46.13

Sánchez, E., Calderón, R., & Herrera, F. (2025). Artificial Intelligence Adoption in SMEs: Survey Based on TOE–DOI Framework, Primary Methodology and Challenges Applied Sciences, 15(12), 6465. https://doi.org/10.3390/app15126465

Nyathani, R., (2023). AI in Performance Management: Redefining Performance Appraisals in the Digital Age Journal of Artificial Intelligence & Cloud Computing, 1-5. https://doi.org/10.47363/jaicc/2023(2)134

Gaurav Gupta, (2025). The Impact of Artificial Intelligence on Modern Program Management International Journal of Scientific Research in Computer Science, Engineering and Information Technology, 11(1), 592-600. https://doi.org/10.32628/cseit25111266

Kovari, A. (2024). AI for Decision Support: Balancing Accuracy, Transparency, and Trust Across Sectors Information, 15(11), 725. https://doi.org/10.3390/info15110725

Oluwaseun Badmus, Shahab Anas Rajput, John Babatope Arogundade, & Mosope Williams, (2024). AI-driven business analytics and decision making World Journal of Advanced Research and Reviews, 24(1), 616-633. https://doi.org/10.30574/wjarr.2024.24.1.3093

Kelly, A. (2024). Impact of Artificial Intelligence on Supply Chain Optimization Journal of Technology and Systems, 6(6), 15-27. https://doi.org/10.47941/jts.2153

, Palade, M., Carutasu, G., (2023). "Organizational Readiness for Artificial Intelligence Adoption" Scientific Bulletin of the Politehnica University of Timişoara Transactions on Engineering and Management, 7(1-2), 30-35. https://doi.org/10.59168/fdms6321

Ifeanyi Onyedika Ekemezie, & Wags Numoipiri Digitemie, (2024). BEST PRACTICES IN STRATEGIC PROJECT MANAGEMENT ACROSS MULTINATIONAL CORPORATIONS: A GLOBAL PERSPECTIVE ON SUCCESS FACTORS AND CHALLENGES International Journal of Management & Entrepreneurship Research, 6(3), 795-805. https://doi.org/10.51594/ijmer.v6i3.936

Felemban, H., Sohail, M., & Ruikar, K. (2024). Exploring the Readiness of Organisations to Adopt Artificial Intelligence Buildings, 14(8), 2460. https://doi.org/10.3390/buildings14082460

Murire, O. T. (2024). Artificial Intelligence and Its Role in Shaping Organizational Work Practices and Culture Administrative Sciences, 14(12), 316. https://doi.org/10.3390/admsci14120316

Ivanova, S., Kuznetsov, A., Zverev, R., & Rada, A. (2023). Artificial Intelligence Methods for the Construction and Management of Buildings Sensors, 23(21), 8740. https://doi.org/10.3390/s23218740

Rane, N., Choudhary, S. P., & Rane, J. (2024). Acceptance of artificial intelligence: key factors, challenges, and implementation strategies Journal of Applied Artificial Intelligence, 5(2), 50-70. https://doi.org/10.48185/jaai.v5i2.1017

Bamidele Micheal Omowole, Amarachi Queen Olufemi-Phillips, Onyeka Chrisanctus Ofodile, Nsisong Louis Eyo-Udo, & Somto Emmanuel Ewim, (2024). Barriers and drivers of digital transformation in SMEs: A conceptual analysis International Journal of Scholarly Research in Science and Technology, 5(2), 019-036. https://doi.org/10.56781/ijsrst.2024.5.2.0037

SCHÖNBERGER, M. (2023). ARTIFICIAL INTELLIGENCE FOR SMALL AND MEDIUM-SIZED ENTERPRISES: IDENTIFYING KEY APPLICATIONS AND CHALLENGES Journal of Business Management, 2189-112. https://doi.org/10.32025/jbm23004

Julien Kiesse Bahangulu, & Louis Owusu-Berko, (2025). Algorithmic bias, data ethics, and governance: Ensuring fairness, transparency and compliance in AI-powered business analytics applications World Journal of Advanced Research and Reviews, 25(2), 1746-1763. https://doi.org/10.30574/wjarr.2025.25.2.0571

Ghosh, U. K. (2025). Transformative AI Applications in Business Decision-Making Advances in Logistics, Operations, and Management Science, 1-40. https://doi.org/10.4018/979-8-3373-1687-1.ch001

Richey, R. G., Chowdhury, S., DavisSramek, B., Giannakis, M., & Dwivedi, Y. K. (2023). Artificial intelligence in logistics and supply chain management: A primer and roadmap for research Journal of Business Logistics, 44(4), 532-549. https://doi.org/10.1111/jbl.12364

Temitayo Oluwaseun Abrahams, Oluwatoyin Ajoke Farayola, Simon Kaggwa, Prisca Ugomma Uwaoma, Azeez Olanipekun Hassan, & Samuel Onimisi Dawodu, (2024). CYBERSECURITY AWARENESS AND EDUCATION PROGRAMS: A REVIEW OF EMPLOYEE ENGAGEMENT AND ACCOUNTABILITY Computer Science & IT Research Journal, 5(1), 100-119. https://doi.org/10.51594/csitrj.v5i1.708

The Impact of Artificial Intelligence on Business Strategy and Decision-Making Processes. (2023). European Economic Letters, https://doi.org/10.52783/eel.v13i3.386

Artificial Intelligence in Project Management: Enhancing Efficiency and Decision-Making. (2024). GLOBAL MAINSTREAM JOURNAL, 1(1), 1-6. https://doi.org/10.62304/ijmisds.v1i1.107

Zahaib Nabeel, M. (2024). AI-Enhanced Project Management Systems for Optimizing Resource Allocation and Risk Mitigation Asian Journal of Multidisciplinary Research & Review, 5(5), 53-91. https://doi.org/10.55662/ajmrr.2024.5502

Celestin, M., & N. Vanitha. (2017). THE SURPRISING ROLE OF AI IN REVOLUTIONIZINGPROJECT MANAGEMENT. ResearchGate, 384–390. https://www.researchgate.net/publication/385700419_THE_SURPRISING_ROLE_OF_AI_IN_REVOLUTIONIZINGPROJECT_MANAGEMENT

Nabeel, M. Z. (2024). AI-Enhanced Project Management Systems for Optimizing Resource Allocation and Risk Mitigation. Asian Journal of Multidisciplinary Research & Review, 5(5), 53–91. https://doi.org/10.55662/ajmrr.2024.5502

Abdul-Azeez, O. Y., Ihechere, A. O., & Idemudia, C. (2024, July 6). Enhancing business performance: The role of data-driven analytics in strategic decision-making. ResearchGate; Fair East Publishers. https://www.researchgate.net/publication/382053812_Enhancing_business_performance_The_role_of_data-driven_analytics_in_strategic_decision-making

Adesina, A. A., Iyelolu, T. V., & Paul, P. O. (2024, June 30). Leveraging Predictive Analytics for Strategic decision-making: Enhancing Business Performance through... ResearchGate; GSC Online Press. https://www.researchgate.net/publication/381900026_Leveraging_predictive_analytics_for_strategic_decision-making_Enhancing_business_performance_through_data-driven_insights

Ahmed, R. (2023). Project performance measures and metrics framework. Edward Elgar Publishing EBooks, 11–22. https://doi.org/10.4337/9781802207613.00007

Damawan, A., & Azizah, S. (2020). Resistance to Change: Causes and Strategies as an Organizational Challenge. ResearchGate. https://www.researchgate.net/publication/339190336_Resistance_to_Change_Causes_and_Strategies_as_an_Organizational_Challenge

Khahro, S. H., Shaikh, H. H., Zainun, N. Y., Sultan, B., & Khahro, Q. H. (2023). Delay in Decision-Making Affecting Construction Projects: A Sustainable Decision-Making Model for Mega Projects. Sustainability, 15(7), 5872. MDPI. https://doi.org/10.3390/su15075872

Medida, L. H., & Kumar, N. (2024). Addressing challenges in data analytics. Advances in Computer and Electrical Engineering Book Series, 16–33. https://doi.org/10.4018/979-8-3693-2260-4.ch002

Molina, A., & Gregson, G. (2017). Real-time evaluation methodology as learning instrument in high-technology SME support networks. International Journal of Entrepreneurship and Innovation Management, 2(1), 69. https://doi.org/10.1504/ijeim.2002.000476

Mortaji, S. T. H., & Shateri, S. (2024). The Role of Data Science in Enhancing Project Management Practices: A Case Study in the Pharmaceutical Industry. Journal of Data Analytics, 3(1), 1–12. https://doi.org/10.59615/jda.3.1.1

Muftah, M. A. R. A. (2022). The Impact of Artificial Intelligence on Auditing Practices and Financial Reporting Accuracy. Integrated Journal for Research in Arts and Humanities, 2(1), 40–46. https://doi.org/10.55544/ijrah.2.1.49

Odejide, O. A., & Edunjobi, T. E. (2024). AI IN PROJECT MANAGEMENT: EXPLORING THEORETICAL MODELS FOR DECISION-MAKING AND RISK MANAGEMENT. Engineering Science & Technology Journal, 5(3), 1072–1085. https://doi.org/10.51594/estj.v5i3.959

Ofer Zwikael, & Jack R. Meredith. (2019, July). (PDF) Evaluating the Success of a Project and the Performance of Its Leaders. ResearchGate. https://www.researchgate.net/publication/334685883_Evaluating_the_Success_of_a_Project_and_the_Performance_of_Its_Leaders

Ogunbukola, M. (2024, September 23). The impact of artificial intelligence on project management: Enhancing efficiency, risk mitigation, and decision-making in complex projects. ResearchGate. https://www.researchgate.net/publication/384266056_The_Impact_of_Artificial_Intelligence_on_Project_Management_Enhancing_Efficiency_Risk_Mitigation_and_Decision-Making_in_Complex_Projects

Omokhafe, S., Durodola, L., None Godbless Ocran, Eweala, J., Echere, Z., & None Adedamola Hadassah Paul-Adeleye. (2024). Challenges and opportunities in AI and digital transformation for SMEs: A cross-continental perspective. World Journal of Advanced Research and Reviews, 23(3), 668–678. https://doi.org/10.30574/wjarr.2024.23.3.2511

Pina, E., Ramos, J., Jorge, H., Váz, P., Silva, J., Wanzeller, C., Abbasi, M., & Martins, P. (2024). Data Privacy and Ethical Considerations in Database Management. Journal of Cybersecurity and Privacy, 4(3), 494–517. MDPI. https://doi.org/10.3390/jcp4030024

Wolniak, R., & Grebski, W. (2023). The concept of diagnostic analytics. Scientific Papers of Silesian University of Technology Organization and Management Series, 2023(175). https://doi.org/10.29119/1641-3466.2023.175.41

Project Introduction

(Nadarajah et al., 2024) illustrates that in today's rapidly evolving digital landscape, cybersecurity has become more critical. In support of this (Ray, 2025) stated, as cyber threats grow in frequency, traditional defense mechanisms often struggle to keep pace with sophisticated attack techniques. Emerging threats reports by CloudFlare, revealed that in the first quarter of year 2025, Cloudflare's system blocked around 20.5 million DDoS attacks, which is a 358% increase year- over -year(Yoachimik & Pacheco, 2025).   According to (Suri babu Nuthalapati, 2023), this alarming trend underscores the urgent need for innovative solutions capable of proactively identifying and mitigating attacks. Integrating machine learning techniques into cybersecurity systems offers a promising pathway to enhance threat detection capabilities, enabling systems to analyze vast amounts of data in real-time, uncover hidden patterns, and predict potential attacks before they cause significant harm (Naima Aamri, 2024). (Aliyu Enemosah & Edmund, 2025) illustrates that by leveraging advanced algorithms, these AI-driven systems can adapt dynamically to evolving threats, creating a more resilient and intelligent defense mechanism. This project aims to explore how machine learning can be seamlessly incorporated into cybersecurity frameworks to develop smarter, adaptive, and more effective protection for digital assets.

THE PROBLEM STATEMENT

Despite cybersecurity evolving, many systems still rely on signature-based detection, which often fails to catch new or zero-day threats (Parameshwar Reddy Kothamali & Banik, 2022). This results in missed attacks and typically a high rate of false positive results that can be difficult to measure, which ultimately harms security efforts. The number of cyber threats is constantly growing in volume and sophistication, and traditional security methods are rapidly becoming obsolete (Rayhan, 2024). This provides a need for intelligent, machine learning solutions that are built to learn from data and improve dynamically. The development of these systems has challenges around data quality, interpretability, and scalability. The goal is an accurate and adaptive threat detection system that can dynamically provide complete protection, fewer false positive results, and less workload for security teams.

MY INSPIRATION FOR THE PROJECT

The motivation behind this project is curiosity about the possibility of using artificial intelligence to revolutionize cybersecurity. After observing the ways cyber attackers constantly evolve, I became fascinated with the idea of ​​developing defense systems that can learn and change in real-time, just like a human expert (Ahsan et al., 2022). The concept of using machine learning and technology to prioritise information to identify trends, and analyse/examine multiple data points to anticipate threats before they appear, is particularly exciting to me. I firmly believe that integrating machine learning into cybersecurity threats will allow us to develop better defenses that are mostly proactive rather than reactive, this would also include the ability to identify patterns in data to anticipate problems and allow the prevention of invisible threats. This project is my way of contributing to a field that has the potential to future-proof our systems by protecting against harmful cyber activities, especially when cyber threats evolve at an unprecedented rate. I chose this project because I want to learn and develop skills that relate to my goals for a career in cybersecurity, and I also expect to gain experience that will be applicable to the cybersecurity field.

A SHORT INTRODUCTION TO THIS PROPOSAL

This proposal outlines a plan to develop and evaluate an integrated cybersecurity framework that leverages machine learning techniques for enhanced threat detection. The focus is on designing models that can analyze network traffic, user behavior, and system logs to identify malicious activities with high precision. The project aims to compare different machine learning algorithms, such as supervised and unsupervised learning, to determine which approaches are most effective in real-world scenarios. Additionally, it will explore how these intelligent systems can be optimized for deployment in AI-driven cybersecurity environments, ensuring they are scalable, interpretable, and capable of evolving with emerging threats. Ultimately, this research seeks to contribute to the development of more adaptive, efficient, and human-like cybersecurity defenses that can keep pace with the rapidly changing digital threat landscape.

This proposal outlines how to develop and evaluate an integrated cybersecurity framework that will utilize machine learning methods to develop a model for detecting threats. The model would learn from the analysis of network traffic, users' behaviors, and system logs to be able to continue to identify malicious activity with a high degree of accuracy. The intent of the project is to conduct comparisons of various supervised and unsupervised machine-learning algorithms. It also examines the intelligent systems and how they might be applied in real-world situations and used in AI-based cybersecurity scenarios, including many factors like scalability, interpretability, and adaptation. Ultimately, this research, if successful, may provide a next-generation approach to adaptive, efficient, human-like cybersecurity that can evolve and be used to protect states from a multitude of threats in an evolving digital landscape.

Next, the proposal will outline the detailed methodology for data collection, model development, and evaluation processes. It will also discuss system integration, testing plans, and potential challenges to ensure a comprehensive understanding of the project.

PROJECT GOAL

The primary objective of this project is to design a machine learning-based framework that improves accuracy, effectiveness, and efficiency in threat detection in AI-enabled cybersecurity systems. The intention is to develop an adaptive model that will facilitate the detection of both known and emerging cyber threats in real-time, enhancing overall security posture.

The secondary goal will be to assess and compare different machine learning algorithms, specifically supervised, unsupervised, and reinforcement learning to examine which techniques are most suitable for the specific types of cyberthreats. This will help to optimize detection and ensure the system remains scalable and robust in variable cybersecurity landscapes.

PROJECT REQUIREMENTS

1. Core project requirement 

• To study and identify appropriate machine learning algorithms (e.g., Random Forests, Support Vector Machines, and Neural Networks) that are effective for this threat detection research to analyze datasets on cybersecurity.

• To obtain and process information on cybersecurity-related data, such as network log data, system alerts, and attack definitions, in order to create training and testing datasets of quality for analysis.

• To develop a threat detection module that combines machine learning models into a single system to process and analyse incoming datasets for real-time decision making.

• To train the machine learning models using labeled datasets to identify patterns associated with normal activity and various cyber threats such as malware, phishing, and intrusion attempts.

• To optimize the models through hyperparameter tuning and feature selection techniques to improve detection accuracy, reduce false positives, and minimize false negatives.

• To evaluate the performance of each model using metrics such as precision, recall, F1-score, and ROC-AUC, and compare their effectiveness in different threat scenarios.

• To implement cross-validation and testing procedures to ensure the robustness and generalizability of the models across diverse datasets.

• To develop a dashboard or reporting interface that visualizes detection results, model performance metrics, and real-time alerts for cybersecurity analysts.

• To document the entire development process, including data collection, model training, deployment, and testing, to ensure reproducibility and facilitate future improvements.

2. ADVANCED PROJECT AIMS

• To incorporate ensemble learning techniques and advanced feature engineering methods to further enhance detection accuracy and reduce false alarms.

• To explore and implement anomaly detection and behavioral analysis methods to identify zero-day threats and sophisticated attack patterns beyond signature-based detection.

• To develop a comparative framework to benchmark the performance of different machine learning models against existing state-of-the-art cybersecurity threat detection systems.

• To conduct a comprehensive analysis of the trade-offs between detection performance and computational resource consumption, especially in resource-constrained environments.

• To experiment with hybrid models combining supervised and unsupervised learning approaches to improve the detection of unseen or evolving cyber threats.

• To investigate the impact of different data preprocessing techniques and feature extraction methods on model performance and stability.

• To generate detailed reports and visualizations comparing the detection efficacy, false positive/negative rates, and resource utilization of each model under various simulated network conditions.

• To propose novel configurations, feature sets, or algorithm modifications based on experimental results that could potentially improve detection capability and operational efficiency.

• To formulate best practice guidelines for deploying AI-based threat detection systems tailored to different organizational needs and network architectures.

• To compare the developed system’s performance with current industry standards and academic research to demonstrate its competitiveness and practical applicability.

Secondary Research Aims

• To examine current research on various machine learning techniques employed for threat detection within AI-driven cybersecurity systems, focusing on their methodologies, capabilities, and limitations.

• To analyze how different machine learning approaches are used to identify and classify cyber threats such as malware, phishing attacks, network intrusions, and zero-day exploits, emphasizing their effectiveness in real-world applications.

• To investigate challenges related to data quality, such as data collection, labeling, imbalance, and noise, and how these impact the performance of machine learning models in cybersecurity contexts.

• To explore the integration of anomaly detection and behavioral analysis methods with machine learning to identify sophisticated and previously unseen cyber threats beyond traditional signature-based detection.

• To review recent advancements in developing adaptive and scalable threat detection frameworks that leverage machine learning for continuous learning and evolution over time.

• To assess the role of feature selection, data preprocessing, and dimensionality reduction techniques in enhancing model accuracy and operational efficiency in cybersecurity environments.

• To compare the performance metrics reported in the literature, such as detection accuracy, false positive rates, and response times, for machine learning-based threat detection systems.

• To identify gaps in existing research, particularly in the areas of real-time processing, system interpretability, and deployment challenges, to inform future improvements.

• To synthesize best practices and emerging trends in deploying AI and machine learning techniques for cybersecurity, emphasizing how these methods contribute to more intelligent, proactive defense systems.

Primary Research

Primary research is the process of collecting data from various resources, including network traffic, system logs, and attack signatures. This data will present respective challenges such as duplicates, values that need to be normalized, and categorical values decode,d that the data scientist is going to manage to ensure the data is reliable. It is likely that the data is going to be drawn from known data sets and be managed through dimensionality reduction and/or feature selection techniques, towards developing the best-performing model. Machine learning algorithms will be developed as part of the overall system, such as anomaly detection methods and classification methods. The various components of the system, data ingestion, feature extraction, model training, and real-time threat assessment will be developed. The work will be delineated into phases to allow a thorough evaluation of existing approaches, gathering and pre-processing the data, establishing models and tuning their parameters, assessing models in practice, testing the system, and determining the reports of outcomes. This is all laid out in a phased manner to allow us to explore possible improvements in threat detection capabilities in cybersecurity systems using machine learning approaches.

Project risk and mitigation

Project objectives

Objective 1 – To conduct a thorough review of current machine learning algorithms used in cybersecurity threat detection, understand their detection parameters, and analyze recent research developments to identify suitable techniques for this project.

Objective 2 – To collect relevant cybersecurity datasets, perform data cleaning, normalization, and feature extraction to prepare high-quality data for training and testing machine learning models.

Objective 3 - To use a variety of machine learning models for threat detection, train those models, optimize the hyperparameters, and measure and evaluate the performance of each using accuracy, precision, recall, and F1-score.

Objective 4 - To take the best machine learning model and integrate it into an AI cybersecurity system, implement real-time threat detection, and assess its level of performance in an emulated performance environment.

Objective 5 - To assess the results of the threat detection system, document the methodology and findings, and make recommendations for improvements and future research related to AI-based cybersecurity threat detection.

Lesson learned

• Obtain extensive knowledge of the different machine learning methods; supervised, unsupervised, ensemble, etc, and how to apply those methods to threat detection.

• Design, train and implement a scalable real-time cybersecurity framework that effectively applies machine learning to detect both known and unknown cyber threats with a high degree of accuracy.

• Develop ability in gathering and cleaning cybersecurity data, potentially using feature selection and dimensionality reduction methods for preparing datasets when modelling.  

• Gain the operational aspects of deploying machine learning models into running systems for cybersecurity purposes, including scalability, resource and interpretability issues.

• Develop problem solving and critical thinking skills through technical debugging, data quality assessment, adapting models and developing from the issues associated with actioning the adaptation process.

• Improve planning and documentation skills through organized research, thorough documentation and project research planning, important for a career in either Cybersecurity, AI development or Data Science.   

• Contribute to global efforts in building a smarter, adaptive security assurance framework to address the rapidly developing threat landscape of cyber-attack backers.

• Handle and develop experience, skills, and knowledge in relation to both specific cybersecurity community models as well as broader community development of skills connected to cybersecurity frameworks as applicable to a career in cybersecurity, AI development, data analysis, and systems integration.

References

Parameshwar Reddy Kothamali, & Banik, S. (2022, March 14). Limitations of Signature-Based Threat Detection. Researchgate. https://www.researchgate.net/publication/388494583_Limitations_of_Signature-Based_Threat_Detection

Rayhan, A. (2024). Cybersecurity in the Digital Age: Assessing Threats and Strengthening Defenses. RG. https://doi.org/10.13140/RG.2.2.31480.25607

Afzal, A., Khan, S., Daud, S., & Butt, A. (2023). Addressing the Digital Divide: Access and Use of Technology in Education. ResearchGate; The Rustam Model School and College (Rustam) Mardan. https://www.researchgate.net/publication/371575436_Addressing_the_Digital_Divide_Access_and_Use_of_Technology_in_Education

Bajger, T., Khoshnaw, D., Ali, K. A. A., & Mousa, K. M. (2025). Impact of Digital Transformation on Rehabilitating Higher Education Infrastructure in Conflict‐Affected Settings. European Journal of Education, 60(3). https://doi.org/10.1111/ejed.70151

Carreno, A. M. (2024). An Analytical Review of John Kotter’s Change Leadership Framework: A Modern Approach to Sustainable... ResearchGate. https://www.researchgate.net/publication/384065104_An_Analytical_Review_of_John_Kotter

Drax, K., Clark, R., Chambers, C. D., Munafò, M., & Thompson, J. (2021). A qualitative analysis of stakeholder experiences with Registered Reports Funding Partnerships. Wellcome Open Research, 6, 230. https://doi.org/10.12688/wellcomeopenres.17029.1

Ellen Ernst Kossek, Porter, C. M., Lindsey Mechem Rosokha, Kelly Schwind Wilson, Rupp, D. E., & Law‐Penrose, J. (2024). Advancing work–life supportive contexts for the “haves” and “have nots”: Integrating supervisor training with work–life flexibility to impact exhaustion or engagement. Human Resource Management, 63(3). https://doi.org/10.1002/hrm.22207

Erceg, A. (2018, May). (PDF) Importance of alignment of strategy and project management. ResearchGate. https://www.researchgate.net/publication/325429664_Importance_of_alignment_of_strategy_and_project_management

Fan, Y. (2024). Accountability in Public Organization: A Systematic Literature Review and Future Research Agenda. Public Organization Review. https://doi.org/10.1007/s11115-024-00792-y

Fennessy, K., Billett, S. R., & Ovens, C. A. (2006, January 1). Learning in and through social partnerships. Researchgate. https://www.researchgate.net/publication/29461889_Learning_in_and_through_social_partnerships

Jeremiah, M., & Kabeyi, B. (2018). Evolution of Project Management, Monitoring and Evaluation, with Historical Events and Projects that Have Shaped the Development of Project Management as a Profession. International Journal of Science and Research (IJSR) ResearchGate Impact Factor, 8(12). https://doi.org/10.21275/ART20202078

King Edward VI College, Stourbridge. (2025, June 21). King Edwards vi College. https://www.kedst.ac.uk/

Kreimeia, A. (2025, June 21). Home. EdTech Hub. https://edtechhub.org/

Kushariyadi, K., Wahid, D. A., Albashori, M. F., Rustiawan, I., ARDENNY, A., & Wahyudiyono, W. (2025). Performance Management Based on Key Performance Indicators (KPI) to improve Organizational Effectiveness. Maneggio, 2(1), 90–102. https://doi.org/10.62872/7yx54j15

Maria José Sampaio de Sá. (2019, May). (PDF) Virtual and Face-To-Face Academic Conferences: Comparison and Potentials. ResearchGate. https://www.researchgate.net/publication/333342205_Virtual_and_Face-To-Face_Academic_Conferences_Comparison_and_Potentials

McKeithan, G. K., Rivera, M. O., Mann, L. E., & Mann, L. B. (2021). Strategies to Promote Meaningful Student Engagement in Online Settings. Journal of Education and Training Studies, 9(4), 1. https://doi.org/10.11114/jets.v9i4.5135

McQuaid, R. (2000). (PDF) the Theory of partnership: Why Have partnerships? ResearchGate. https://www.researchgate.net/publication/291300642_The_theory_of_partnership_Why_have_partnerships

Oluwaseun Kayode Akinsola, Adedokun Taofeek, Kingsley Onu, & Yinka Owoeye. (2025, January 31). Legal Challenges and Best Practices for Structuring Corporate Partnerships, Joint Ventures, and Strategic Alliances. Researchgate. https://www.researchgate.net/publication/388653168_Legal_Challenges_and_Best_Practices_for_Structuring_Corporate_Partnerships_Joint_Ventures_and_Strategic_Alliances

Oyeniyi Author. (2025). STRATEGIC MANAGEMENT OF SOFTWARE PROJECTS: COST, RISK, CONTINGENCY, BUDGET AND SCHEDULE. ResearchGate. https://www.researchgate.net/publication/391940931_STRATEGIC_MANAGEMENT_OF_SOFTWARE_PROJECTS_COST_RISK_CONTINGENCY_BUDGET_AND_SCHEDULE

Pandita, A., & Kiran, R. (2023). The Technology Interface and Student Engagement Are Significant Stimuli in Sustainable Student Satisfaction. Sustainability, 15(10), 7923. https://doi.org/10.3390/su15107923

Rafiq, S., Iqbal, S., & Afzal, D. A. (2024, May 21). The Impact of Digital Tools and Online Learning Platforms on Higher Education Learning Outcomes. ResearchGate; unknown. https://www.researchgate.net/publication/380734414_The_Impact_of_Digital_Tools_and_Online_Learning_Platforms_on_Higher_Education_Learning_Outcomes

Reddy, R. (2023, August 1). EFFECTIVE COMMUNICATION STRATEGIES FOR STAKEHOLDER ENGAGEMENT. Researchgate. https://doi.org/10.13140/RG.2.2.13590.36162

Santiago Jr., C. S., Ulanday, M. L. P., Centeno, Z. J. R., Bayla, M. C. D., & Callanta, J. S. (2021). Flexible learning adaptabilities in the new normal: E-learning resources, digital meeting platforms, online learning systems and learning engagement. Asian Journal of Distance Education, 16(2). https://doi.org/10.5281/zenodo.5762474

Sharma, S. (2025). Enhancing Employability through Skill Development: A Study in the Context of NEP 2020. International Journal of Research Publication and Reviews, 6(4), 5143–5147. https://doi.org/10.55248/gengpi.6.0425.1317

Steve Rowlinson. (2007, January). (PDF) Procurement Systems: A cross-industry project management perspective. ResearchGate. https://www.researchgate.net/publication/301813670_Procurement_Systems_A_cross-industry_project_management_perspective

Tahili, M. H., Tolla, I., Ahmad, M. A., Samad, S., Saman, A., & Pattaufi, P. (2022). Developing the strategic collaboration model in basic education. International Journal of Evaluation and Research in Education (IJERE), 11(2), 817. https://doi.org/10.11591/ijere.v11i2.21907

Vining, A. R., & Boardman, A. E. (2024). Cost–benefit analysis and “next best” methods to evaluate the efficiency of social policies: As in pitching horseshoes, closeness matters. Annals of Public and Cooperative Economics. https://doi.org/10.1111/apce.12484

Wheelahan, L. M., Moodie, G., Lavigne, E., & Samji, F. (2018, October). Case Study of TAFE and Public Vocational Education in Australia: Preliminary Report. ResearchGate; unknown. https://www.researchgate.net/publication/334173113_Case_Study_of_TAFE_and_Public_Vocational_Education_in_Australia_Preliminary_Report

Yadav, N. (2024, January 30). The Impact of Digital Learning on Education. ResearchGate; International Surya Publication. https://www.researchgate.net/publication/378177781_The_Impact_of_Digital_Learning_on_Education

Zalsos, E., & Corpuz, G. G. (2024). Academic Management and Instructional Practices of Higher Education Institutions in Lanao Del Norte: Basis for Faculty Development Plan. American Journal of Arts and Human Science, 3(2), 19–38. https://doi.org/10.54536/ajahs.v3i2.2649

Project Introduction

1.1 The Field of Study

In today's digital environment, cloud computing has fundamentally transformed the storage, processing, and management of data across various industries. (Venkata Reddy Keesara, 2025). As (Akinade et al., 2024) indicated that more organizations move to cloud services, ensuring the security of these platforms becomes critically important to address emerging vulnerabilities. In support of this (Ahmad et al., 2021) stated that while cloud platforms offer remarkable flexibility and scalability, they also introduce new risks, especially as threat actors discover additional attack points amidst the rapid deployment of expansive infrastructure with temporary and elastic resources. According to (Qudus, 2025) depending on a combination of security tools to safeguard these large and resilient systems can inadvertently expand an organization’s attack surface. The author (Aggrey et al., 2025) stated that applying effective security measures helps bolster the overall security posture of cloud environments. As per (Dawood et al., 2023), intrusion detection systems (IDS) are a vital component in this regard, playing a strategic role in defending cloud setups. An IDS continuously monitors network traffic and system activities, enabling the quick detection and response to malicious behaviors (Diana, Dini and Paolini, 2025). According to (Chauhan and Shiaeles, 2023), IDSs are positioned at the frontline of cloud security frameworks, helping to identify potential disruptions that could threaten sensitive data or operational.

1.2 The Problem Statement

(Nasim, Pranav and Dutta, 2025) stated that Intrusion detection systems (IDS) play a vital role in safeguarding cloud-based networks, but there’s still a significant knowledge gap regarding which IDS tools are most effective in different cloud scenarios, especially when it comes to handling various types of attacks. As cyber threats grow more sophisticated, a report conducted by Statista indicated that the worldwide annual cost of cybercrime could reach around $15.63 trillion by 2029 (Petrosyan, 2024). The author (Ahmadi, 2024) indicated that there is a strong desire among organizations to invest in IDS solutions that can effectively balance detection, requirements, and response capabilities to address vulnerabilities in the cloud. In support of this, (Sridharan, 2017) stated that choosing the optimal tool is not straightforward, primarily because cloud environments are constantly growing in terms of traffic patterns and requirements.  The author (Tolu Michael, 2025) stated that there is currently no consensus among popular open-source IDS options like Snort, Suricata, and Zeek regarding which is best suited for different cloud network conditions specially under changing traffic loads. According to (Dawood et al., 2023) the lack of a clear standard underscores the necessity for more comprehensive evaluations and comparisons of these tools, so that best practices can be developed for their deployment across diverse cloud security architectures. 

1.3 My Inspiration for the Project

My enthusiasm for networking and cybersecurity encourages me to see how different protective tools behave in real-world environments based on the cloud (Aslan et al., 2023). According to me, understanding and evaluating different IDS tools can be crucial in improving the overall security strategy (Sindika, Nicholaus and Hamadi, 2024). Due to digitalization, a large number of organizations and individuals are moving towards cloud storage. According to a report conducted by IT Security Wire, globally 94% of organizations use cloud computing (Bureau, 2025. Also experiencing a large number of cyber threats. According to AAG, in 2024 there will be about 6 billion cyber-attacks all over the world (Griffiths, 2025). I am motivated to find practical ways for an organization to better protect something so valuable (Cremer et al., 2022). I find that the field of cybersecurity is more demanding and has possibilities for technology and professional growth. I chose this project because I want to develop my skills in cybersecurity and contribute to protecting digital assets. From a personal perspective, I expect to gain practical experience that will directly support my future career in cybersecurity.In this research I can provide valuable information by reporting findings on increasing understanding and implementation of effective security in the cloud (Rao Narendra, Sr Tadapaneni and Sabri, 2020).

1.4 A Short Introduction to This Proposal

The main purpose of this report is to offer a detailed comparison of three most popular open-source intrusion detection systems such as Snort, Suricata, and Zeek within a cloud-based network environment (Bada, Nabare and Quansah, 2020). This project helps to understand how all the three systems perform in terms of detection accuracy, resource usage, and response times under different network conditions. The various network conditions like heavy traffic loads, encrypted data, or changing workloads (El-Hajj, 2025). Testing of all three IDS tools using real data and simulated cyber attacks is designed to replicate real-world scenarios and analyze how these IDS tools work in real operational settings (Dini et al., 2023). The valuable results of this research will help organizations to enhance their security strategies, and also implement these IDS tools to defend against complex threats in cloud environments. Ultimately, the results will provide recommendations on which IDS configurations are most effective in different situations of cyber-attacks. The subsequent sections of this proposal will detail the methodology for testing the IDS tools, including the setup, data collection, and analysis procedures. Additionally, the report will present expected outcomes, potential challenges, and the significance of this research for cybersecurity practices.

Project Goal

This project intends to examine detection accuracy and operational performance of three market leading intrusion detection systems (IDS) Snort, Suricata, and Zeek under different environmental settings. The secondary aim of the project is to compare performance metrics to analyse which is most reliable and resource cognizant for real time network invigilation.

Project Requirement

1. Core project requirement 

• To install three IDS (Snort, Suricata & Zeek), onto a managed network for trial purposes (Day and Burns, 2011).

• To configure all of the IDS to appropriate rules and settings such that they are capable of optimal detection

• To simulate different types of network traffic, and capture normal activity as well as various forms of cyber-attacks through internet traffic generators and datasets

• To monitor and log each performance in real-time, such that the analysis could identify accuracy of detection, response times, and resource consumption i.e., CPU and memory (Li et al., 2019).

• To assess the systems' efficiency in terms of processing speeds and resource consumption, under different network loads and the complexity of traffic

• To aggregate and analyze the value of the data collected, as well as report on the capabilities or weaknesses of each system in each of its scenarios

• To develop reports and visualizations of the metrics reported for each IDS in detail and supplementally in terms of comparisons

• To provide recommendations or configurations that perform better in detection accuracy and operational efficiency

2. Advanced Project aims

• To describe configuration parameters of IDS tools such as Snort, Suricata, and Zeek, as well as tuning parameters, to enable the most accurate detection across many forms of attack, while minimizing false positives and false negatives (Wahyu et al., 2023).

• To build a benchmark framework to measure performance for each IDS in varying network conditions, from network load and high-traffic volume to encrypted traffic.

• To compare the IDS performance, along with their detection metrics, with recent state-of-the-art research on the subject in the academic or commercial sector, while referencing current industry-state-of-the-art research (Alkasassbeh and Al-Haj Baddar, 2022).

• To suggest novel combinations/new configurations or changes to existing rules and algorithms in the IDS realm, based upon findings from experimental analyses that are intended to enhance detection efficacy (Wahyu et al., 2023).

• To create a series of best practice guides and recommendations for deploying and configuring IDS tools for different network environments to enhance security efficacy.

Purpose of Conducting the Literature Review:  

This project uses a quantitative-based experimental methodology to understand current intrusion detection systems, with a special focus on the open-source tools of Snort, Suricata, and Zeek. It analyzes their detection capability, resource usage, response time, performance in different network conditions, such as increased traffic rates or current concerns for encrypted traffic. This knowledge will offer a foundation for designing experiments and mocking evaluation methods for IDS in cloud environments, while allowing for a valid comparison that is substantiated by established performance metrics and best practices.

1. Performance Metrics and Detection Capabilities:  

The identification of each IDS, Snort, Suricata, and Zeek's ability to perform in terms of accuracy in detection, false positive rates, and response time is delineated in the review by referencing previous empirical studies.

2. Resource Utilization and Efficiency:  

Examining the resource consumption (CPU, memory) of each IDS, particularly with high network traffic. Understanding these factors is important in the deployment of IDS in cloud environments, where resource usage can be important.

3. Operational Environments and Configurations:  

The review will analyse how the various configurations, rules set, and deployment will influence Snort, Suricata, and Zeek’s performance. It will also address the flexibility to encrypted traffic and high traffic networks, resulting in the best implementation practices for those interested in cloud.

Goals 

• To get a better understanding of the detection capabilities, resource usage, and response times of Snort, Suricata, and Zeek based on existing literature.

• To understand the limitations and strengths of the different IDS tools in various network situations of high traffic, as well as encrypted data and mobile workloads.

• To gather benchmark and performance ratings for relevant cloud-based settings to compare these tools effectively.

• To use the experiment literature to design the experiment and the performance evaluation will align with the standards in industry and in academia.

Primary Research

The research plan describes the method for assessing and comparing the performance of Snort, Suricata, and Zeek in a variety of network environments (Lankos, 2014). The focus will be on evaluating the operational performance of these intrusion detection systems (IDS) in a cloud-based environment. The research analysis  will take place under different parameters to assess detection accuracy, false positive and false negative rates, response times to security incidents, under varying conditions such as traffic volume and encryption status. The data will be gathered through consistent logging and analysis based on live or simulated sessions involving network traffic. The end goal is to create as standardized a process as possible so the evaluation means are parallel across all three IDS tools. 

The investigation will be carried out in several phases:

Phase 1: Setup and Configuration (2 weeks)  

In this initial  stage, all three IDS (Snort, Suricata, and Zeek) will be installed on virtual machines within a cloud-based network service such as Microsoft Azure. Each IDS will be installed on a separate virtual server with the appropriate network interfaces. In order to simulate an actual deployment and maximize the utility of the IDS, the cloud infrastructure will have virtual servers active for various network services (e.g., web server, FTP server, SSH server). Enough of the network topology will be designed to simulate routing traffic through these services and the IDS, and adhere to industry best practices. The rules will be configured to detect real-time attacks (e.g., DdoS, brute-force, network reconnaissance, and normal traffic flows) respective of each IDS. The proper logging will be in sense to track detailed performance and detection information.

Phase 2: Performance Data Collection (3-4 weeks)  

During this phase, the IDS will undergo testing in the various traffic conditions generated from the virtual servers. The different conditions will simulate expected traffic under normal conditions, attack traffic (including, but not limited to, DDoS, brute-force, and reconnaissance). Traffic will be generated using traffic simulation tools that approximate realistic scenarios. The IDS will be monitored to look at detection accuracy, response times, false positives/negatives, and resource consumption. To ensure clear distinctions, all aspects of each test will remain consistent except for the varying conditions and data will be collected on detection rates, latency, CPU and memory usage, and throughput.

Phase 3: Data Analysis and Comparison (2 weeks)  

The collected data will be statistically analyzed to evaluate the detection capabilities, detection rate, response time, and resource consumption of Snort, Suricata, and Zeek. The rates to be calculated are detection rate, false positive rate, false negative rate, and resource consumption. The analysis allows to  identify which IDS performs better in detecting and responding to various types of attacks under different network traffic scenarios.

Phase 4: Reporting and Benchmarking (2 weeks)  

The final report will synthesize findings of each IDS and compare their performance to other industry and academic benchmarks. The final report will discuss the strengths and weaknesses of each tool and will also provide recommendations for typical environments to deploy tools given results. The final report will also offer recommendations for deploying tools in real-world environments in the cloud based on detection efficacy and resource usage.

Project Resources

Risk and Mitigation

Project Objectives

• To conduct a detailed literature review of existing network intrusion detection systems focused on Snort, Suricata, and Zeek, assessing detection capabilities, performance measures, and the applicability of these tools in different network settings.

• To evaluate and compare the performance of Snort, Suricata, and Zeek in detecting various types of network intrusions, including measurements on detection accuracy, false positive rate, false negative rate, response time, and resource utilization.

• To examine the operational features, strengths, and weaknesses of each IDS tool in real-world circumstances, and assess the capability of these tools depending upon network security requirements.

• To provide a report documenting the methodology used, results obtained, challenges experienced, and conclusions reached to provide a clear comparison against each set of tools.

Reference list

Alkasassbeh, M. and Al-Haj Baddar, S. (2022). Intrusion Detection Systems: A State-of-the-Art Taxonomy and Survey. Arabian Journal for Science and Engineering. doi:https://doi.org/10.1007/s13369-022-07412-1.

Almarshad, F.A., Zakariah, M., Ghada Abdalaziz Gashgari, Eman Abdullah Aldakheel and Abdullah (2023). Detection of Android Malware Using Machine Learning and Siamese Shot Learning Technique for Security. IEEE Access, 11, pp.127697–127714. doi:https://doi.org/10.1109/access.2023.3331739.

Day, D.J. and Burns, B.M. (2011). A Performance Analysis of Snort and Suricata Network Intrusion Detection and Prevention Engines. The Fifth International Conference on Digital Society. [online] Available at: https://www.researchgate.net/publication/241701294_A_Performance_Analysis_of_Snort_and_Suricata_Network_Intrusion_Detection_and_Prevention_Engines.

GeeksforGeeks (2019). Intrusion Detection System (IDS) - GeeksforGeeks. [online] GeeksforGeeks. Available at: https://www.geeksforgeeks.org/intrusion-detection-system-ids/.

Gupta, A. and Sharma, L.S. (2019). Performance Evaluation of Snort and Suricata Intrusion Detection Systems on Ubuntu Server. Lecture Notes in Electrical Engineering, pp.811–821. doi:https://doi.org/10.1007/978-3-030-29407-6_58.

Kaspersky (2021). Kaspersky for Business Machine Learning for Malware Detection. [online] Available at: https://media.kaspersky.com/en/enterprise-security/Kaspersky-Lab-Whitepaper-Machine-Learning.pdf.

Langkos, S. (2014). Research Methodology: Data collection method and Research tools. [online] ResearchGate. Available at: https://www.researchgate.net/publication/270956555_CHAPTER_3_-_RESEARCH_METHODOLOGY_Data_collection_method_and_Research_tools.

Li, S., Jia, Z., Li, Y., Liao, X., Xu, E., Liu, X., He, H. and Gao, L. (2019). Detecting Performance Bottlenecks Guided by Resource Usage. IEEE access, 7, pp.117839–117849. doi:https://doi.org/10.1109/access.2019.2936599.

Martin, D. and Ailab (2011). Evaluation: From precision, recall and F-measure to ROC, informedness, markedness & correlation. [online] ResearchGate. Available at: https://www.researchgate.net/publication/276412348_Evaluation_From_precision_recall_and_F-measure_to_ROC_informedness_markedness_correlation.

None S.V. Vasantha, Kiranmai, N.B., Jyoshna, N.B., Ramakrishna, N.S., Suvarnamukhi, N.B. and Hariharan, N.S. (2024). Detecting Malware on Android Devices Using CNN-LSTM with FDSWM based Feature Selection. Nanotechnology Perceptions. [online] doi:https://doi.org/10.62441/nano-ntp.vi.1391.

Prof. Shashikant V Golande, Sanket Vaidya, Aniket Pardeshi, Vivekanand Katkade and Vedant Pawar (2024). An Efficient Network Intrusion Detection and Classification System using Machine Learning. International Journal of Advanced Research in Science, Communication and Technology, pp.267–272. doi:https://doi.org/10.48175/ijarsct-22045.

Sabbah, A., Taweel, A. and Zein, S. (2023). Android Malware Detection: A Literature Review. Communications in Computer and Information Science, pp.263–278. doi:https://doi.org/10.1007/978-981-99-0272-9_18.

Vergara Cobos, E. and Cakir, S. (2024). A Review of the Economic Costs of Cyber Incidents. [online] Available at: https://documents1.worldbank.org/curated/en/099092324164536687/pdf/P17876919ffee4079180e81701969ad0a18.pdf.

Wahyu, A.P., Fauziah, K., Nahrowi, A.S., Faiz, M.N. and Muhammad, A.W. (2023). Strengthening Network Security: Evaluation of Intrusion Detection and Prevention Systems Tools in Networking Systems. ProQuest. [online] doi:https://doi.org/10.14569/IJACSA.2023.0140934.

YANG, J., TANG, J., YAN, R. and XIANG, T. (2022). Android Malware Detection Method Based on Permission Complement and API Calls. Chinese Journal of Electronics, 31(4), pp.773–785. doi:https://doi.org/10.1049/cje.2020.00.217.

Introduction

The global food industry is navigating a period of profound transformation, shaped by rising consumer expectations, urgent sustainability goals, and accelerated technological disruption. (Statista, 2024) reported that the global food market is expected to reach US$9.43tn in 2025. The market is expected to grow annually by 6.41% (CAGR 2025-2030), digital innovation and eco-conscious practices have moved from being competitive advantages to business essentials. Further, Australia is no exception to this reality(Amin-Chaudhry, Young and Afshari, 2022). A study conducted by (Statista, 2022) reported that, australia food industry is a critical contributor to the economy, expected to generate US$96.59 billion in revenue in 2025, growing at an annual rate of 4.82%. Absolutely, it can help smooth out the transitions and enhance the overall flow: The Australian food sector is navigating a complex landscape shaped by both global pressures and local challenges, calling for adaptive and forward-thinking management strategies. In response, the Australian government has committed to reducing food wastage by 50% by 2030 (National Retail Association, 2023), signaling a national push toward sustainability. Complementing this initiative, major restaurants such as Dishroom , dominos, Starbucks, California pizza kitchen etc are leveraging AI-powered systems to reduce inventory costs by up to 20%, illustrating a growing reliance on technology to drive operational efficiency (Estes, 2024). 

Concurrently, shifting consumer preferences particularly the 22% year-on-year rise in Australians adopting plant-based protein diets (Khara, Riedy & Ruby, 2021) are prompting firms to rethink production strategies to align with emerging health and environmental values.These interrelated trends government policy, technological innovation, and evolving consumer behavior are collectively reshaping the strategic priorities of the sector. Rather than isolated developments, they represent a broader transformation in the way food businesses operate and compete. This study, through a qualitative thematic literature review, explores the key drivers of this shift, evaluates the current state of adaptation within the industry, and offers strategic recommendations for firms aiming to build resilience and maintain competitiveness in a post-pandemic, climate-conscious economy. In doing so, it emphasizes the necessity for businesses to move beyond reactive approaches and proactively engage with emerging trends to secure long-term success.

Research Aim

To investigate transformative trends in commerce and management in Australia’s food industry (2018–2025) through thematic literature analysis to inform strategic adaptation and resilience to thrive in the competitive market.

Research Objectives

1. To identify and critically examine the key technological, environmental, and consumer-driven trends that have influenced commerce and management strategies in the Australian food sector.
2. To assess how food businesses in Australia have adapted their operational practices, supply chains, and decision-making processes in response to these emerging trends.
3. To evaluate the effectiveness of current strategies in enhancing business resilience, sustainability, and market competitiveness within a rapidly evolving industry landscape.

• To propose evidence-based strategic recommendations for strengthening the future adaptability and sustainability of Australian food businesses amid ongoing and projected challenges.

Research Questions

1. What emerging trends have significantly influenced commerce and management strategies in Australia’s food sector from 2018 to 2025, and how have businesses adapted their operations and decision-making in response to technological, environmental, and consumer-driven changes?
2. What strategic recommendations can be made to enhance the resilience, sustainability, and competitiveness of Australian food businesses in the face of ongoing and future challenges?

Literature review

According to (YouMatter, 2019), globalization has traditionally been defined as the increasing interconnectedness of economies, societies, and cultures through trade, investment, technology, and migration. (Fujitsu Global, 2020) illustrates that it was once synonymous with physical mobility, global supply chains, and centralized production. However, (Zaidan, Cochrane & Belal, 2025) illustrates that the COVID-19 pandemic has fundamentally altered this definition, shifting the focus from physical integration to digital connectivity, regional resilience, and decentralized operations. (Gilbert, 2023) demonstrates that even before the pandemic, globalization came under pressure due to rising protectionism, trade wars, and growing political nationalism. (Siripurapu, Berman & Fong, 2024) indicated that events like Brexit and US-China trade tensions signaled a move toward economic insularity. (Altman, 2020) highlighted the pandemic accelerated these trends, causing the largest and fastest decline in international flows including trade, FDI, and travel in modern history.

Yet, (Augustinraj, 2017) indicates that globalization did not end. While physical flows declined, digital globalization surged(View, 2017). For instance, Zoom’s daily meeting participants skyrocketed from 10 million in December 2019 to over 300 million by April 2020 (Statista, 2022), illustrating how businesses adapted to remote collaboration and virtual operations. This shift has redefined globalization: today, one can work across borders without ever leaving home. The (IMD Business School, 2023) emphasizes that globalization is not disappearing but evolving. Companies are adopting “glocal” strategies balancing global efficiency with local responsiveness(Andrés Fernández Miguel et al., 2024). Apple’s diversification of its manufacturing base beyond China, including expansions in India and Vietnam, reflects this shift (Vengattil, 2025).

From a business perspective, this new form of globalization demands greater agility, digital infrastructure, and supply chain resilience(Li, Chen and Guo, 2025). Organizations must now manage operations in a world where remote work, automation, and regionalization are the norm (Tiwari, Sharma & Jha, 2024). According to (Deloitte Insights, 2025), global GDP growth will depend heavily on how businesses and governments adapt to these new realities. In essence, globalization today is less about physical presence and more about digital reach(Globalization , n.d.). Businesses must rethink their models not just to survive but to thrive in a world where location is no longer a limitation, and resilience is the new competitive edge(Pham et al., 2021).

Research Gap:

While existing literature effectively traces the transformation of globalization from physical interconnectedness to a digitally driven model, several gaps remain underexplored. Most current studies such as (Yin and Ran, 2022) focus on macroeconomic trends and corporate adaptations, such as supply chain diversification and remote collaboration tools. However, there is limited empirical research on the long-term impacts of digital globalization on small and medium enterprises (SMEs), labor dynamics, and workforce resilience, especially in developing economies (surajwancy and kulkarni, 2024). Moreover, although the concept of “glocalization” is gaining traction, further investigation is needed into how businesses can sustainably balance global efficiency with local responsiveness in volatile geopolitical and public health contexts (Ghemawat, 2009). There is also a gap in understanding how digital infrastructure disparities across regions affect equitable participation in the new globalization model (Ho, Ngoc and Ho, 2025). Hence, while scholars agree that globalization is evolving rather than disappearing, future research must delve into the nuanced consequences of this shift, particularly its implications for inclusivity, economic equity, and organizational adaptability at different scales.

Methodology

This study will adopt a qualitative literature review framework, utilizing manual thematic analysis and coding to investigate emerging trends in commerce and management within Australia’s food industry. This method will facilitate a rich, interpretive synthesis of academic and industry literature while allowing the researcher to stay closely engaged with the material throughout the analytic process.

1. Research Design
   A thematic analysis approach will be selected to uncover recurring concepts and narratives across the literature. Manual coding will allow for direct engagement with texts, encouraging a deeper reflexive interpretation than would be achieved through automated software (Williams, 2024).

2. Data Sources and Selection Criteria
   Literature will be collected from scholarly databases such as Google Scholar, Direct Science, and Web of Science, alongside credible industry and government publications (e.g., ABARES, Food Innovation Australia Limited). Sources will be selected based on relevance, credibility, and alignment with the study’s focus, namely, trends in AI adoption, sustainability, consumer behavior, and supply chain management in Australia's food sector. Only English-language publications from 2018 to 2025 will be included.

Ethical Considerations
Despite being a literature-focused study and excluding primary human participants, ethical integrity will remain paramount throughout this study. This review will rely on the following principles:

• Transparency and Reflexivity: Thematic coding will be conducted manually. The researcher will maintain a record of every analytic decision, reflection, and interpretation. This will ensure transparency and help limit personal bias during the coding and development of themes (DELVE, 2024).

• Credible Representation of Ideas: The study will aim to faithfully represent the literature, with the intention of making interpretations of authors' perspectives honest and avoiding any misrepresentation of their ideas. The researcher will read the literature multiple times in an effort to retain the nuance and motivation present in original works (O’Connor & Joffe, 2020).

Potential Outcomes

This research aims to offer a comprehensive understanding of how commerce and management practices in Australia’s food industry have evolved in response to technological, environmental, and consumer-driven pressures (Kent et al., 2022). It will identify key trends from 2018 to 2025 that have shaped strategic decisions, operational changes, and innovation among food businesses of all sizes.A major focus will be on small and medium-sized enterprises (SMEs), highlighting how they’ve adapted despite limited resources. The study will explore how these firms are using affordable digital tools, adopting sustainable practices, and sourcing locally to overcome supply chain challenges. It will assess their adoption of AI and e-commerce to boost competitiveness and sustainability.In light of ongoing disruptions like COVID-19, climate change, and global economic instability, the research will evaluate the effectiveness of resilience strategies, operational agility, and market positioning. It will pinpoint best practices and innovations that support adaptability and cost management.The final outcome will be evidence-based recommendations tailored to SMEs, offering practical pathways to build resilience, promote sustainability, and enhance competitiveness. These insights aim to empower smaller food businesses to thrive amid ongoing industry change and continue contributing to Australia’s economy.

Research Plan

Research will take place over a six-month period. After the topic is approved, literature review and manual coding will occur with universally ongoing design of the thematic analysis methodology. Data collection (sourcing and reviewing literature) will begin and will inform theme development, followed by the structured analysis and writing of findings, and final revisions and submission at the end of the timeline.

Timeline Table

Albattat, A., Fakir, F.Z., Yi, Z. and Hussain, Z. (2025). Innovative Trends Shaping Food Marketing and Consumption. [online] ResearchGate. doi:https://doi.org/10.4018/979-8-3693-8542-5.

Nikunj (2025). Top Reasons Why Restaurants Fail Within First Year Of Operations. [online] Restroworks Blog. Available at: https://www.restroworks.com/blog/top-reasons-why-restaurants-fail-within-the-first-year-of-operations/.

Statista (2022). Food - Australia | Statista Market Forecast. [online] Statista. Available at: https://www.statista.com/outlook/cmo/food/australia.

Statista (2024). Food - Worldwide | Statista Market Forecast. [online] Statista. Available at: https://www.statista.com/outlook/cmo/food/worldwide.

Dwivedi, Y.K., Ismagilova, E., Hughes, D.L. and Carlson, J. (2021). Setting the future of digital and social media marketing research: Perspectives and research propositions. International Journal of Information Management, [online] 59(1), pp.1–37. doi:https://doi.org/10.1016/j.ijinfomgt.2020.102168.

Food AI: A game changer for Australia’s (2024). Food AI: A game changer for Australia’s F&B sector | FaBA. [online] Faba.au. Available at: https://faba.au/news/food-ai-a-game-changer/.

Guo, Y., Liu, F., Song, J.-S. and Wang, S. (2024). Supply Chain Resilience: a Review from the Inventory Management Perspective. Fundamental Research, [online] 5(2), pp.1–14. doi:https://doi.org/10.1016/j.fmre.2024.08.002.

Szenderák, J., Fróna, D. and Rákos, M. (2022). Consumer Acceptance of Plant-Based Meat Substitutes: A Narrative Review. Foods, [online] 11(9), p.1274. doi:https://doi.org/10.3390/foods11091274.

Amin-Chaudhry, A., Young, S. and Afshari, L. (2022). Sustainability motivations and challenges in the Australian agribusiness. Journal of Cleaner Production, [online] 361(1), p.132229. doi:https://doi.org/10.1016/j.jclepro.2022.132229.

Introduction and Rationale

According to (Alkharafi & Alsabah, 2025), globalization has traditionally been viewed as the process of increasing interconnectedness among nations through trade, investment, technology, and cultural exchange, fostering economic growth and development. (Panwar, Pinkse and Marchi, 2022) demonstrates that before the COVID-19 pandemic, this interconnectedness was evident in expansive global supply chains, international labor mobility, and the free flow of goods, capital, and ideas across borders. (Bappa Grace Hosen, 2023) revealed that in such an environment, businesses thrived, borderless trade and seamless international cooperation were widely considered as essential drivers of market expansion and global competitiveness. However, the onset of the COVID-19 crisis profoundly disrupted this narrative(Mockaitis, Butler and Ojo, 2022). As (Xu et al., 2020) emphasize, the pandemic paralyzed global supply chains, restricted movement, and exposed fundamental vulnerabilities in traditional globalization models. According to (UNCTAD, 2021), global merchandise trade volumes dropped by 5.3% in 2020, and foreign direct investment flows dropped by 35% highlighting the fragility of overextended global networks. These disruptions have prompted a rethinking of globalization's core principles. (McKinsey, 2022) indicates the focus has shifted toward resilience, regionalization, and digital connectivity as key pillars of a more adaptable global system. Businesses are now increasingly adopting “glocal” strategies that combine global efficiency with local responsiveness (IMD, 2025). For example, (U.S. Companies , 2025) illustrates that multinational firms are diversifying supplier bases, near-shoring production, and investing in digital infrastructure to build more robust and flexible operations. This transformation is not only strategic, it is measurable. The report conducted (Cross-border data flows, 2024) shows that cross-border data flows have increased by over 45% since 2019, signaling a major acceleration of digital globalization. Meanwhile, (Liu et al., 2025) elaborates that regional trade agreements, such as the Regional Comprehensive Economic Partnership (RCEP), are gaining prominence, reshaping how countries collaborate economically. In light of these shifts, the definition of globalization is evolving. It is no longer solely about borderless integration, but increasingly about building adaptive, regional, and digitally connected networks capable of withstanding global shocks(Monteiro and Barata, 2025). This evolving model has profound implications for how companies operate, compete, and strategize in today's dynamic environment. This study aims to explore these changing dimensions of globalization, assess their impact on business strategies, and evaluate whether globalization remains a key driver of market development in the post-pandemic world. For policymakers, multinational corporations, and global stakeholders, understanding this transformation is critical to navigating the complexities of a new era in global economic integration.

Aim 

The aim of the project is to  critically analyze the transformation of globalization in the aftermath of the COVID-19 pandemic and evaluate its continuing relevance and influence as a key driver of global markets.

Objective 

1. To identify how global market dynamics have shifted in response to pandemic-driven disruptions.

2. To evaluate whether globalization remains a central mechanism for economic growth in the post-pandemic world.

3. To explore emerging trends such as regionalism, digital globalization, and supply chain resilience.

4. To assess policy responses and strategies from key global economies in shaping the future of globalization.

5. To propose policy and strategic recommendations for governments and businesses on how to navigate globalization's evolving role, including diversification of supply chains, investing in digital infrastructure, and fostering international cooperation for future resilience.

Research Question

Is globalization still a critical driver of global market dynamics in the post-COVID-19 era, or has its influence been reshaped by emerging geopolitical, technological, and economic trends?

Brief Literature Review

According to (YouMatter, 2019), globalization has traditionally been defined as the increasing interconnectedness of economies, societies, and cultures through trade, investment, technology, and migration. (Fujitsu Global, 2020) illustrates that it was once synonymous with physical mobility, global supply chains, and centralized production. However, (Zaidan, Cochrane & Belal, 2025) illustrates that the COVID-19 pandemic has fundamentally altered this definition, shifting the focus from physical integration to digital connectivity, regional resilience, and decentralized operations. (Gilbert, 2023) demonstrates that even before the pandemic, globalization came under pressure due to rising protectionism, trade wars, and growing political nationalism. (Siripurapu, Berman & Fong, 2024) indicated that events like Brexit and US-China trade tensions signaled a move toward economic insularity. (Altman, 2020) highlighted the pandemic accelerated these trends, causing the largest and fastest decline in international flows including trade, FDI, and travel in modern history.

Yet, (Augustinraj, 2017) indicates that globalization did not end. While physical flows declined, digital globalization surged(View, 2017). For instance, Zoom’s daily meeting participants skyrocketed from 10 million in December 2019 to over 300 million by April 2020 (Statista, 2022), illustrating how businesses adapted to remote collaboration and virtual operations. This shift has redefined globalization: today, one can work across borders without ever leaving home. The (IMD Business School, 2023) emphasizes that globalization is not disappearing but evolving. Companies are adopting “glocal” strategies balancing global efficiency with local responsiveness(Andrés Fernández Miguel et al., 2024). Apple’s diversification of its manufacturing base beyond China, including expansions in India and Vietnam, reflects this shift (Vengattil, 2025).

From a business perspective, this new form of globalization demands greater agility, digital infrastructure, and supply chain resilience(Li, Chen and Guo, 2025). Organizations must now manage operations in a world where remote work, automation, and regionalization are the norm (Tiwari, Sharma & Jha, 2024). According to (Deloitte Insights, 2025), global GDP growth will depend heavily on how businesses and governments adapt to these new realities. In essence, globalization today is less about physical presence and more about digital reach(Globalization , n.d.). Businesses must rethink their models not just to survive but to thrive in a world where location is no longer a limitation, and resilience is the new competitive edge(Pham et al., 2021).

Research Gap

While existing literature effectively traces the transformation of globalization from physical interconnectedness to a digitally driven model, several gaps remain underexplored. Most current studies such as (Yin and Ran, 2022) focus on macroeconomic trends and corporate adaptations, such as supply chain diversification and remote collaboration tools. However, there is limited empirical research on the long-term impacts of digital globalization on small and medium enterprises (SMEs), labor dynamics, and workforce resilience, especially in developing economies (surajwancy and kulkarni, 2024). Moreover, although the concept of “glocalization” is gaining traction, further investigation is needed into how businesses can sustainably balance global efficiency with local responsiveness in volatile geopolitical and public health contexts (Ghemawat, 2009). There is also a gap in understanding how digital infrastructure disparities across regions affect equitable participation in the new globalization model (Ho, Ngoc and Ho, 2025). Hence, while scholars agree that globalization is evolving rather than disappearing, future research must delve into the nuanced consequences of this shift, particularly its implications for inclusivity, economic equity, and organizational adaptability at different scales.

Research Methodology

This study adopts a qualitative research approach centered on a systematic literature review to explore the evolving nature and definition of globalization in the post-COVID-19 era, with a particular focus on its impact on business operations and organizational strategies. The research investigates how globalization has shifted from traditional models of physical integration and cross-border trade to more digitally enabled, decentralized, and resilient systems.

Data Collection Plan

The study relies on secondary data analysis, drawing from academic journals, institutional reports, and professional publications. Sources will include peer-reviewed journals (e.g., Global Policy, International Business Review), global institutions (e.g., World Bank, IMF, WTO), and consulting firms (e.g., McKinsey, Deloitte). Data will be accessed through online databases such as Google Scholar, ResearchGate, ScienceDirect, and IEEE Xplore.

Search terms will include:

• “Post-COVID globalization”

• “Remote work and global collaboration”

• “Digital trade and virtual economies”

• “FDI decline and regionalization”

• “Supply chain resilience”

• “Geopolitical shifts and economic nationalism”

Search string: (Globalization AND Evaluation) AND (Covid-19 OR Coronavirus) AND ("Market Driver" OR "Economic Impact")

The literature will be limited to works published from 2019 onward, ensuring relevance to the post-pandemic global landscape.

Sampling Strategy

As the study does not involve human participants, a purposeful sampling strategy will be used to select literature that reflects diverse regional perspectives and credible insights. Priority will be given to peer-reviewed research and institutional reports that align with the study’s objectives.

Data Analysis

The study will employ manual thematic analysis, a qualitative method for identifying and interpreting patterns across the literature. This approach is well-suited to understanding the complex, fast-evolving discourse around globalization’s transformation.

Initial coding will identify recurring terms such as:

• “Remote work,” “digital platforms,” and “virtual collaboration” → grouped under “Digital Globalization”

• “Reshoring,” “regional supply chains,” and “economic nationalism” → grouped under “Geopolitical Realignment”

• “Automation,” “AI integration,” and “resilience” → grouped under “Structural Shifts in Business Models”

These themes will help the study:

1. Conceptual Shifts: Compare globalization definitions before and after COVID-19.

2. Scholarly Perspectives: Identify key contradictions and agreements in academic debates.

3. Strategic Responses: Examine how businesses adapted to shifting global dynamics.

Potential Outcomes

The study anticipates uncovering that globalization, though disrupted by the COVID-19 crisis, has not diminished but significantly transformed. It is expected to reveal a shift from traditional, physically integrated models to more digitally enabled and regionally resilient systems. Businesses are increasingly adopting hybrid strategies that blend global reach with local agility, reshaping competitive dynamics. Digital globalization, characterized by remote work, virtual collaboration, and rising data flows, is emerging as a central pillar of the new economic landscape. Regional trade agreements and nearshoring practices further indicate a trend toward decentralized yet interconnected economies. The research will likely reinforce the idea that globalization remains a vital market driver, though now propelled more by technology and strategic adaptation than by physical integration. This evolution demands that policymakers and corporations reassess their approaches, emphasizing digital infrastructure, flexibility, and supply chain resilience to thrive in an unpredictable global environment shaped by future disruptions.

Research Plan / Timeline

Reference list

Andrés Fernández Miguel, Maria Pia Riccardi, García-Muiña, F.E., Fernández, A.P., Valerio Veglio and Davide Settembre-Blundo (2024). From Global to Glocal: Digital Transformation for Reshoring More Agile, Resilient, and Sustainable Supply Chains. Sustainability, 16(3), pp.1196–1196. doi:https://doi.org/10.3390/su16031196.

Augustinraj, A.B. (2017). Globalization is not ending, it’s changing. [online] mint. Available at: https://www.livemint.com/Opinion/XF7o6jevNo1OBOpu4DGZ1I/Globalization-is-not-ending-its-changing.html.

Bappa Grace Hosen (2023). Navigating the Borderless Horizon: A Review Study of Challenges & Opportunities of Borderless World. ResearchGate, [online] 8(2), pp.33–41. Available at:https://www.researchgate.net/publication/381962793_Navigating_the_Borderless_Horizon_A_Review_Study_of_Challenges_Opportunities_of_Borderless_World [Accessed 29 May 2025].

Cross-border data flows (2024). Cross-border data flows. [online] OECD. Available at: https://www.oecd.org/en/topics/sub-issues/cross-border-data-flows.html.

Fujitsu Global. (2020). Achieving Efficiency and Resilience Across Global Supply Chains with Digital Technology. [online] Available at: https://www.fujitsu.com/global/vision/insights/20-supplychain/.

Globalization (n.d.). Globalization is becoming more about data and less about stuff | McKinsey. [online] www.mckinsey.com. Available at: https://www.mckinsey.com/mgi/overview/in-the-news/globalization-is-becoming-more-about-data-and-less-about-stuff.

Li, P., Chen, Y. and Guo, X. (2025). Digital Transformation and Supply Chain Resilience. International Review of Economics & Finance, 99(104033), p.104033. doi:https://doi.org/10.1016/j.iref.2025.104033.

Liu, C., Zhou, J., Wen, W., Liu, F. and Zhang, C. (2025). The Effect of the Regional Comprehensive Economic Partnership on Taiwan’s Global Value Chain of the Electronic Information Industry. Sustainability, [online] 17(1), p.281. doi:https://doi.org/10.3390/su17010281.

McKinsey (2022). Building supply chain resilience, but risks remain | McKinsey. [online] www.mckinsey.com. Available at: https://www.mckinsey.com/capabilities/operations/our-insights/taking-the-pulse-of-shifting-supply-chains.

Mockaitis, A.I., Butler, C.L. and Ojo, A. (2022). COVID-19 pandemic disruptions to working lives: A multilevel examination of impacts across career stages. Journal of Vocational Behavior, 138, p.103768. doi:https://doi.org/10.1016/j.jvb.2022.103768.

Monteiro, J. and Barata, J. (2025). Digital Twin-Enabled Regional Food Supply Chain: A Review and Research Agenda. Journal of Industrial Information Integration, [online] p.100851. doi:https://doi.org/10.1016/j.jii.2025.100851.

Panwar, R., Pinkse, J. and Marchi, V.D. (2022). The Future of Global Supply Chains in a Post-COVID-19 World. California Management Review, 64(2). doi:https://doi.org/10.1177/00081256211073355.

Pham, L.D.Q., Coles, T., Ritchie, B.W. and Wang, J. (2021). Building business resilience to external shocks: Conceptualising the role of social networks to small tourism & hospitality businesses. Journal of Hospitality and Tourism Management, 48, pp.210–219. doi:https://doi.org/10.1016/j.jhtm.2021.06.012.

U.S. Companies (2025). Will More U.S. Companies Turn to Nearshoring to Diversify Supply Chains? | Inbound Logistics. [online] Inbound Logistics. Available at: https://www.inboundlogistics.com/articles/will-more-u-s-companies-turn-to-nearshoring-to-diversify-supply-chains-in-the-next-five-years/.

UNCTAD (2021). Global merchandise trade exceeds pre-COVID-19 level, but services recovery falls short | UNCTAD. [online] unctad.org. Available at: https://unctad.org/news/global-merchandise-trade-exceeds-pre-covid-19-level-services-recovery-falls-short.

View (2017). Economics of Digital Globalization and Information Data Flows. [online] Viewpoints which Matter. Available at: https://chaturvedimayank.wordpress.com/2017/02/26/economics-of-digital-globalization-and-information-data-flows/ [Accessed 23 Jun. 2025].

Project Introduction

According to (Mohamed, 2025), the field of Artificial Intelligence (AI) has become increasingly vital in addressing modern cybersecurity challenges, particularly in the detection and prevention of sophisticated digital threats. Currently, email is one of the widely used communication tools, however, its security has become a critical concern these days  both for individuals and organizations (Altulaihan et al., 2023). Further, (Nadeem et al., 2023) demonstrates that email security prones to a number of security issues, but phishing is one of the most prominent threats, where malicious actors follow the guise of trusted entities. As per (Enforcement OF DIRECTORATE, n.d.), phishing was the most reported cybercrime for the fifth consecutive year, accounting for over 800,000 complaints and causing estimated global losses exceeding $3.5 billion. As a result, platforms such as Microsoft Defender for Office 365, have adopted advanced AI Models that block more than 1 million phishing emails per day with 99.9998% true positives (as of April 2025) (Melanie_Cohen, 2025). Further, (Saswata Dey, Writuraj Sarma & Sundar Tiwari, 2023) continues to explain how many AI-powered systems such as DeepPhish utilize ML, deep learning, and NLP to analyze the contents of email, sender behavior, and embedded links in real-time. Although these advances highlight the role of AI in the prevention of phishing, they also highlight the limitations of existing solutions (Naseer, 2024). 

However, despite these technological strides, existing models often face challenges in generalizing across diverse phishing strategies and adapting to evolving attack patterns (S. Kavya & D. Sumathi, 2024). Many systems rely heavily on static rules or historical data, which can be easily bypassed by more dynamic and obfuscated phishing attempts. Additionally, false positives remain a concern, potentially disrupting legitimate communications (Dalalah & Dalalah, 2023). Therefore, there is an urgent need to develop more agile, context-aware ML models that continuously learn and respond to emerging phishing tactics in real time.

Hence, this project explores the development and evaluation of a realtime phishing detection system using machine learning, aiming to enhance email security through intelligent automation and adaptive threat mitigation. By leveraging AI’s ability to learn from evolving attack patterns, the proposed system seeks to provide a more robust and proactive defense against phishing threats in modern communication environments (Velusamy et al. n.d.).

Inspiration for the Project

This project was inspired by a deep rooted interest in the transformative potential of Artificial Intelligence in the field of cybersecurity. With the increase of data breaches, email scams, and the sophistication of phishing, I was intrigued with how I could use machine learning as more than just a predictive tool, but a form of real-time defence (Denys Spys 2025). My education, as well as my independent research, has often converged on the intersection of intelligent systems and digital safety: how operators could utilize machines that had been trained to recognize patterns, adapt to threats, and help protect users from evermore deceptive cyber tactics. As someone who is looking for a career path developing AIdriven security solutions, I see this project as a worthwhile step towards gaining relevant and meaningful hands-on experience in this field, which is constantly evolving.

A Short Introduction to This Proposal

This proposal outlines the development and evaluation of a realtime phishing detection system powered by machine learning algorithms (Adeyemi Onih 2024). This begins with a statement of the project aim and objectives,followed up with a detailed list of the basic and advanced functionalities that will be built. The document also contains the analysis and discussion of your answer, a description of the chosen research methodology, the project resources and references, a risk management plan, completion dates, and a project closure procedure. More generally, the document will provide a plan, or unique design, for realizing a practical research-based solution in the domain of AI based cybersecurity.

Project Goal or Aim

Primary aim 

The primary aim of this project is to develop a realtime phishing detection system using machine learning to enhance the security of AI-driven email platforms.

Secondary aim

The secondary aim is to compare the performance of two selected ML algorithms to recommend the most effective model for phishing detection.

Project Requirements and Functionalities

Core Project Requirements

• To collect a balanced and labeled dataset of phishing and legitimate emails, including metadata, header information, and body content for training purposes.

• To select and implement two supervised machine learning algorithmsRandom Forest and Support Vector Machine (SVM)for phishing detection.

• To engineer and extract relevant features from the dataset, such as URL patterns, sender behavior, keyword frequency, and hyperlink structure.

• To preprocess and normalize the dataset using techniques such as tokenization, stopword removal, and TFIDF for accurate textbased analysis.

• To split the data into training, validation, and testing sets using appropriate techniques such as crossvalidation to avoid overfitting.

• To train the preferred algorithms using the pre-prepared datasets and to finetune hyperparameters for best performance.

• To test out the trained models against metrics of accuracy, precision, recall, F1score, and ROCAUC, with particular attention given to restricting false positives and false negatives.

• To develop and implement a modular Python based framework that incorporates the trained model into a realtime email scanning pipeline for inference and alerts.

• To test the validity of the system in a simulated environment that replicates real world email traffic, confirming robustness and responsiveness under unpredictable movements.

Advanced Project Aims

• To enhance detection accuracy by integrating ensemble methods (e.g., stacking classifiers) or incorporating deep learning techniques such as LSTM or BERT for semantic-level analysis (if time permits).

• To compare the model’s results against state of the art benchmarks reported in the literature and analyze discrepancies or improvements in detection rate and generalization.

• To establish real-time detection functionality by developing an asynchronous message extractor or queuebased architecture approach (simulating RabbitMQ or Kafka) to simulate an enterprise-level email stream.

• To perform error analysis of the misclassified emails with emphasis on feature weight and confusion matrix trends for refinement.

• To record and consider the ethical implications of automated email analysis, such as user privacy, bias, and false accusation potential.

• To present recommendations for placing the model into a production enterprise email security system, with emphasis on scalability, security, and maintenance.

Secondary Research Aims (Literature Review)

The secondary research aims of this project are:

• To identify and select the most suitable machine learning algorithms (e.g., Random Forest, Support Vector Machine, CNN, LSTM) for realtime phishing detection based on their reported accuracy, precision, recall, and F1scores in recent academic and industry studies.

• To examine which emailbased features (e.g., URL structure, sender information, patterns in the subject line, nested links) commonly exist and are effective in phishing detection models, and construct and apply these features to the feature engineering. 

• To obtain and examine performance results from state of the art phishing detection systems so as to compare and contrast with the outcomes of this project, and to rationalize the decisions made regarding the algorithms and evaluation metrics used in the implementation. (Noronha et al. 2022).

Primary Research

This project will adopt a datadriven, experimental approach to develop and evaluate a realtime phishing detection system using machine learning. The primary research activity will incorporate dataset acquisition, preprocessing, model development, and ROI evaluation in a structured phase approach.

Dataset Collection and Preprocessing

A public dataset, labeled (phishing and legitimate), will be gathered from sites such as Kaggle, PhishTank, UCI Machine Learning Repository. The dataset will feature email metadata (sender address, subject line), feature variables based on the content body (text, URLs embedded), and email headers.

Preprocessing steps will include:

• Parsing and cleaning raw email data to extract structured fields.

• Text normalization (lowercasing, punctuation removal, stopword filtering).

• Feature engineering, including:

• URLbased features (length, presence of IPs, special characters).

• Content-based features (keyword frequency, HTML tags).

• Metadata features (sender domain reputation, timestamp patterns).

• Vectorization using TFIDF or word embeddings for textual features.

• Label encoding and data balancing using techniques like SMOTE if class imbalance is detected.

System Development (Artefact)

The core artefact will be a Python-based phishing detection system that integrates:

• A feature extraction pipeline for realtime email analysis.

• Two supervised ML models (e.g., Random Forest and Support Vector Machine) trained on the preprocessed dataset.

• A modular architecture allowing for easy model swapping and future scalability.

• A realtime inference module simulating email flow and flagging suspicious messages.

Phases of Investigative Work

1. Data Acquisition & Preprocessing (Week 1–2)

• Identify and download dataset(s).

• Clean, normalize, and engineer features.

• Split data into training, validation, and test sets.

2. Model Implementation & Training (Week 3–4)

• Implement Random Forest and SVM classifiers.

• Train models using crossvalidation and hyperparameter tuning.

3. Testing & Evaluation (Week 5)

• Evaluate models on unseen test data using metrics such as:

• Accuracy

• Precision

• Recall

• F1score

• ROCAUC

• Analyze false positives and false negatives.

4. Benchmarking & Comparison (Week 6)

• Compare results with state of the art benchmarks identified in the literature review.

• Justify algorithm selection based on empirical evidence.

5. System Integration & Simulation (Week 7)

• Integrate the trained model into a simulated email environment.

• Test realtime detection capabilities and latency.

6. Documentation & Final Review (Week 8)

• Compile findings, document methodology, and reflect on limitations and future improvements.

Project Resources

Project Risks and Their Mitigation

Project Plan

Project Objectives

• Objective 1: To undertake a comprehensive literature review of ML-based phishing detection systems, analyze algorithms used, evaluate their reported accuracy, and determine relevant detection features.

• Objective 2: Identify and obtain an appropriate phishing dataset, then clean, preprocess, and extract features from it . 

• Objective 3: Implement two ML models (e.g., Random Forest and SVM), train the processed dataset, and conduct initial performance evaluations .

Project Outcomes and Lessons to Be Learned

• Gain a solid understanding of how ML models can be applied to cybersecurity applications, particularly phishing detection.

• Develop the ability to preprocess and engineer features from real-world email datasets for model training.

• Acquire practical experience comparing ML algorithms and finetuning them for realtime detection performance.

• Recognize the difficulties in finding high precision and high recall in imbalanced classification tasks. 

• Hone critical thinking skills by reading academic literature and relating it to actual implementation results.

• Enhance research skills, time management, technical report writing, and presenting.

• Build the foundation for a potential future career in AI based cybersecurity, model evaluation, or threat detection systems.

Reference list

Adeyemi Onih, V 2024, ‘Phishing Detection Using Machine Learning: A Model Development and Integration’, International Journal of Scientific and Management Research, vol. 07, no. 04, pp. 27–63.

Altulaihan, EA, Alismail, A, Rahman, MMH & Ibrahim, AA 2023, ‘Email Security Issues, Tools, and Techniques Used in Investigation’, Email Security Issues, Tools, and Techniques Used in Investigation, vol. 15, no. 13, pp. 10612–10612.

Denys Spys 2025, Phishing Statistics in 2025: The Ultimate Insight | TechMagic, Blog | TechMagic .

Ekramul Haque Tusher, Ismail, MA, Rahman, MA, Alenezi, AH & Uddin, M 2024, ‘Email Spam: A Comprehensive Review of Optimize Detection Methods, Challenges, and Open Research Problems’, IEEE Access, Institute of Electrical and Electronics Engineers, pp. 1–1.

Enforcement OF DIRECTORATE n.d., viewed 21 June 2025, <https://enforcementdirectorate.gov.in/sites/default/files/2025-05/Annual_Report_24-25.pdf>.

Eze, CS & Shamir, L 2024, ‘Analysis and Prevention of AI-Based Phishing Email Attacks’, Electronics, vol. 13, no. 10, p. 1839.

Folasade.Yetunde Ayankoya, Olasubomi Priscilla Olakunle, Daniel Uchechukwu Umezurike & Chioma Favour Ekpetere 2025, ‘Enhancing organizational cybersecurity: A framework for mitigating email phishing attacks’, Global Journal of Engineering and Technology Advances, vol. 23, GSC Online Press, no. 3, pp. 038–047, viewed 21 June 2025, <https://gjeta.com/sites/default/files/GJETA-2025-0169.pdf>.

Melanie_Cohen 2025, Microsoft Defender for Office 365’s Language AI for Phish: Enhancing Email Security, TECHCOMMUNITY.MICROSOFT.COM, viewed 21 June 2025, <https://techcommunity.microsoft.com/blog/microsoftdefenderforoffice365blog/microsoft-defender-for-office-365s-language-ai-for-phish-enhancing-email-securit/4410446>.

Minocha, S & Singh, B 2022, ‘A novel phishing detection system using binary modified equilibrium optimizer for feature selection’, Computers & Electrical Engineering, vol. 98, p. 107689.

Mohamed, N 2025, ‘Artificial intelligence and machine learning in cybersecurity: a deep dive into state-of-the-art techniques and future paradigms’, Knowledge and Information Systems, Springer Science and Business Media LLC.

Nadeem, M, Syeda Wajiha Zahra, Muhammad Noman Abbasi & Ahmed, W 2023, Phishing Attack, Its Detections and Prevention Techniques, ResearchGate, Springer Nature.

Naseer, I 2024, ‘The role of artificial intelligence in detecting and preventing cyber and phishing attacks’, European Journal of Engineering Science and Technology, vol. Vol. 11, Mokslines leidybos deimantas, MB, no. 9, pp. 82–86.

Noronha, MDM, Henriques, R, Madeira, SC & Zárate, LE 2022, ‘Impact of metrics on biclustering solution and quality: A review’, Pattern Recognition, vol. 127, Pergamon, p. 108612.

Obaloluwa Ogundairo & broklyn, peter 2024, ‘AI-Driven Phishing Detection Systems’, Journal of Cyber Security, Tech Science Press.

Osamor, J, Ashawa, M, Alireza Shahrabi, Philip, A & Iwendi, C 2025, ‘The Evolution of Phishing and Future Directions: A Review’, International Conference on Cyber Warfare and Security, vol. 20, no. 1, pp. 361–368.

Saswata Dey, Writuraj Sarma & Sundar Tiwari 2023, ‘AI-powered phishing detection: Integrating natural language processing and deep learning for email security’, World Journal of Advanced Engineering Technology and Sciences, vol. 10, no. 2, pp. 394–415.

Tsehay Admassu Assegie, Ayodeji Olalekan Salau, Chhabra, G, Kaushik, K & Sepiribo Lucky Braide 2024, ‘Evaluation of Random Forest and Support Vector Machine Models in Educational Data Mining’, vol. 2019, pp. 131–135.

Velusamy, Y, Thilagavathi, C, Rehna, B, Joseph, Rajeswari, M & Sai, J n.d., ‘Intelligent Phishing Detection and Mitigation Framework Using Advanced AI Techniques’, vol. 24, p. 2025, viewed 21 June 2025, <https://erode-sengunthar.ac.in/wp-content/uploads/2025/05/1.Saveetha.pdf>.

S. Kavya, & D. Sumathi. (2024). Staying ahead of phishers: a review of recent advances and emerging methodologies in phishing detection. Artificial Intelligence Review, 58(2). https://doi.org/10.1007/s10462-024-11055-z

Dalalah, D., & Dalalah, O. M. A. (2023). The false positives and false negatives of generative AI detection tools in education and academic research: The case of ChatGPT. The International Journal of Management Education, 21(2), 100822. https://doi.org/10.1016/j.ijme.2023.100822

Project Introduction

The rapid expansion of the ecommerce industry has transformed the way consumers shop online, with global online retail sales reaching approximately $5.2 trillion in 2021 and projected to grow to over $8.2 trillion by 2026 (Mahajan 2024). This growth has led to an increasingly competitive landscape, where personalized shopping experiences are key to attracting and retaining customers. Companies that nail personalization generate 40% more revenue, according to a McKinsey report that also noted high levels of customer satisfaction (McKinsey 2021)

Recommendation systems based on a machine learning (ML) algorithm are key to delivering these personalized experiences. These systems capture and analyze largescale datasets, such as browsing history, purchasing patterns, and user preferences to generate personalized product recommendations (Akash Takyar 2023). For example, Amazon’s recommendation engine is responsible for 35% of its sales, demonstrating the significant impact of effective personalization (Gonçalvesá & Pinheiro 2023). As a result, recommendation systems contribute to improved customer engagement, conversion, and brand loyalty in an increasingly competitive digital environment.

Nevertheless, (Yi, Kim & Ju 2022) reveals that as customer demographics become progressively more diverse, traditional recommendation approaches do not always provide relevant recommendations based on demographics for different customer segments. (McKinsey 2021b) shows that, approximately 71% of consumers expect companies to proactively anticipate their needs and preferences, yet most of the recommendation systems lack demographic considerations which prevent small individualization not large scale personalization. For instance, preferences can vary widely based on age, gender, and cultural background; Generation Z, for example, tends to prioritize sustainability and trendiness, whereas older consumers may focus more on quality and durability(Salam et al. 2024).

Addressing this challenge,(Nabil, Chkouri & Bouhdidi 2024) highlight the importance of incorporating demographic based information such as age into recommendation algorithms. Incorporating such as data that can substantially increase the relevance of the recommendation, thereby resulting in greater satisfaction, conversion rates, etc. According to a study (Yin, Qiu & Wang 2025), personalized recommendations based on age increased clickthrough rates by up to 20-30%.

Inspiration of the project

The inspiration behind this project stems from the growing need for smarter, more inclusive recommendation systems that cater to a diverse customer base. By integrating demographic insights into machine learning models, businesses can refine their recommendations, ensuring they resonate more deeply with individual consumers leading to enhanced engagement and increased sales. Beyond its commercial impact, this project aligns with my passion for AI, machine learning, and data science. It presents an opportunity to explore innovative solutions in the field of ecommerce, pushing the boundaries of personalization and predictive analytics. Moreover, working on this initiative will contribute to the development of critical skills that will shape my future career as an AI developer or data analyst, positioning me at the forefront of cutting-edge advancements in intelligent systems.

A short introduction to this proposal

The ecommerce industry is evolving rapidly, with global online retail sales projected to surpass $8.2 trillion by 2026 (Mahajan 2024). As competition intensifies, personalized shopping experiences have become essential for customer engagement and retention. Recommendation systems powered by machine learning play a crucial role in this personalization, analyzing user behavior to deliver tailored product suggestions (Akash Takyar 2023). However, traditional recommendation models often overlook demographic factors, limiting their effectiveness for diverse consumer groups. This project aims to address this gap by integrating demographic insights such as age and gender into machine learning-based recommendation algorithms with clear objectives, well-defined technical stages, evaluation metrics, and advanced modeling strategies that collectively structure the project into aims, requirements, development, and outcomes. By enhancing personalization at an individual level, businesses can significantly boost customer satisfaction, conversion rates, and overall revenue. Beyond its commercial impact, this initiative aligns with the growing need for inclusive AI solutions and presents an opportunity to push the boundaries of predictive analytics. It also serves as a foundation for developing essential skills in AI and data science, preparing for a future at the forefront of intelligent systems innovation.

Project Goal or Aim

Primary aim

The primary aim of this project is to develop a personalized product recommendation system for ecommerce platforms that leverages collaborative filtering techniques, with a specific focus on improving recommendation accuracy for different age groups.

Secondary aim

To evaluate the impact of age-integrated collaborative filtering on user engagement metrics such as clickthrough rate, dwell time, and repeat purchases across distinct age segments.

Core Project Requirements

2.1 Core Project Requirements

• Data Collection and Preprocessing
• Collect publicly available ecommerce datasets that include user ratings, demographic information (age, gender), purchase history, and browsing behavior.
• Perform data cleaning to handle missing, inconsistent, or noisy data through imputation and validation techniques.
• Normalize and encode demographic features (e.g., age groups, gender categories) to enable their integration into the recommendation system.
• Split the dataset into training and testing subsets, ensuring representative demographic distribution in both.
• Exploratory Data Analysis (EDA)
• Analyze the distribution of user demographics and purchase patterns to identify key trends and potential biases.
• Visualize correlations between demographic attributes and product preferences to inform feature engineering.
• Design and Implementation of DemographicAware Collaborative Filtering Algorithms
• Implement userbased collaborative filtering, incorporating demographic data to weight user similarity metrics (e.g., adjusted cosine similarity, Pearson correlation).
• Implement itembased collaborative filtering, integrating demographic insights to improve item similarity calculations.
• Develop hybrid models that combine demographic features with traditional collaborative filtering to enhance recommendation relevance.
• Optimize similarity measures and hyperparameters (e.g., neighborhood size, similarity thresholds) through grid search or other tuning methods.
• Model Training and Validation
• Train the recommendation algorithms on the prepared training dataset.
• Validate model performance using metrics such as Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Precision, Recall, and F1score.
• Incorporate demographic segmentation to evaluate how well recommendations perform across different age groups and genders.
• Evaluation and Performance Analysis
• Compare demographic-aware models against baseline collaborative filtering models that do not consider demographic data.
• Analyze the impact of demographic features on recommendation accuracy and user satisfaction.
• Measure the system's scalability and computational efficiency, especially with larger datasets.
• User Interface and Usage Scenario Simulation
• Develop a simple interface (e.g., Jupyter Notebook dashboards) to simulate user interactions and demonstrate personalized recommendations based on demographic profiles.
• Test the recommendation system with various demographic inputs to assess relevance and diversity.
• Documentation and Reporting
• Document the system architecture, algorithms used, and experimental procedures.
• Present performance metrics, insights, and limitations.
• Provide recommendations for realworld implementation and potential improvements.

2.2 Advanced Project Aims

• Achieve Maximum Possible Recommendation Accuracy
• Finetune hybrid demographic-aware collaborative filtering models to reach the highest achievable precision and recall across all user segments.
• Investigate advanced similarity measures, such as cosine similarity with demographic weighting, or matrix factorization techniques that incorporate demographic features.
• Comparison with StateoftheArt Techniques
• Benchmark the developed system against recent research studies and industry standard recommendation frameworks that incorporate demographic data.
• Evaluate the improvements in personalization, engagement, and conversion metrics relative to existing models.
• Hybrid Model Development
• Develop more sophisticated hybrid models combining collaborative filtering, content-based filtering, and demographic features to address coldstart and sparsity issues.
• Implement deep learning-based approaches (e.g., neural collaborative filtering) that leverage demographic embeddings for enhanced personalization.

Secondary Research Aims (Literature Review)

• Evaluating the algorithms employed in existing demographic-aware recommendation systems, focusing on their accuracy, efficiency, and effectiveness across diverse user segments.
• Identifying current limitations such as data quality issues, scalability challenges, and personalization gaps that hinder system performance.
• Gathering performance metrics and findings from recent research to compare with my project’s approach and inform optimal algorithm and feature choices.

Primary Research

• Data Collection & Preprocessing: Acquire publicly available ecommerce datasets with user ratings, demographics, and purchase history; clean, encode, normalize data; split into training and testing sets ensuring demographic balance.
• Model Development & Evaluation: Implement demographic-aware collaborative filtering algorithms; train and optimize models; evaluate performance using metrics like RMSE, MAE, Precision, Recall, and analyze results across different demographic segments.
• Research Phases & Timeline: Conduct data analysis (Weeks 1-2), develop algorithms (Weeks 3-4), train and tune models (Weeks 5-6), test and validate (Weeks 7-8), demonstrate system and document findings (Weeks 9-10)

Project Resources

Project risks and their mitigation

Project Outcomes and Lessons to Be Learned

• Development of a personalized recommendation system that effectively incorporates age demographics to improve relevance(Nadeem and Suleman, 2024).
• Enhanced understanding of collaborative filtering techniques and their application in demographic-aware systems(Εμμανουήλ Βοζαλής et al., 2004b).
• Comparative insights into userbased versus itembased filtering approaches within an ecommerce context.
• Hansen has experience in data collection, preprocessing, model training, and hyperparameter tuning.
• Lessons in dealing with the difficulties of demographic data integration in recommender systems and methods of tackling these issues.
• Personal development in research, data analysis, and systems building; experience the best way to be positioned for future career opportunities in AI and data driven solutions.

Reference list

Akash Takyar 2023, How to Build an AIpowered Recommendation System, LeewayHertz  Software Development Company.

GonçalvesSá, J & Pinheiro, F 2023, ‘Societal Implications of Recommendation Systems: A Technical Perspective’, Law, governance and technology series, Springer International Publishing, pp. 47–63.

Mahajan, A 2024, 5 Killer Strategies to Enhance eCommerce Revenue in 2024, sphinx solution , viewed 12 June 2025, <https://www.sphinxsolution.com/blog/increaseecommercerevenue/>.

McKinsey 2021, The Value of Getting Personalization rightor wrongis Multiplying, McKinsey & Company.

Nabil, S, Chkouri, MY & Bouhdidi, JE 2024, ‘Demographic information combined with collaborative filtering for an efficient recommendation system’, International Journal of Electrical and Computer Engineering (IJECE), vol. 14, no. 5, p. 5916.

Salam, KN, Singkeruang, AWTF, Husni, MF, Baharuddin, B & A.R, DP 2024, ‘GenZ Marketing Strategies: Understanding Consumer Preferences and Building Sustainable Relationships’, Golden Ratio of Mapping Idea and Literature Format, vol. 4, no. 1, pp. 53–77.

Yi, S, Kim, D & Ju, J 2022, ‘Recommendation technologies and consumption diversity: An experimental study on product recommendations, consumer search, and sales diversity’, Technological Forecasting and Social Change, vol. 178, p. 121486.

Yin, J, Qiu, X & Wang, Y 2025, ‘The Impact of AIPersonalized Recommendations on Clicking Intentions: Evidence from Chinese ECommerce’, Journal of theoretical and applied electronic commerce research, vol. 20, Multidisciplinary Digital Publishing Institute, no. 1, pp. 21–21.