Authors: Mr. R. Nagarajan
Abstract: In today's competitive market environment, business analysis has developed into a necessity beyond just describing reports to becoming a strategy for data-driven decision making. This research paper offers an integrated approach to applying business analysis to managerial decision analytics for creating a sustainable competitive advantage in the e-commerce setting. This approach utilizes the weighted bipartite graph network model, in combination with graph neural networks (GNN), that takes into account recency, profit, and priority parameters in modeling the customer-product relationship. Quantitative analysis of transactions data through the use of the above-mentioned approach shows remarkable results in the following aspects: customer retention predictions (18.5%), inventory optimization (23.7%), and dynamic pricing (31.2%). Comparative analysis shows that graph-based models surpass the performance of traditional machine learning approaches by 12-15% in terms of predictive accuracy.
DOI: http://doi.org/10.5281/zenodo.22724924
