Synergistic Neural Matrix Factorization: Elevating Complementary Product Recommendations in E-Commerce using Deep Neural Networks

This research explores the submission of deep neural networks (DNNs) to enhance corresponding product references in e-commerce stages. By leveraging the Neural Collaborative Filtering (NeuMF) model, which integrates Comprehensive Matrix Factorization (GMF) and Multi-Layer Perceptron (MLP), the study captures both linear and non-linear user-item connections to improve recommendation accuracy. Key performance metrics, including accuracy, loss, ROC curve, and precision-recall curve, were analyzed to assess the model's effectiveness. Results show that the NeuMF model significantly improves recommendation accuracy and user satisfaction, with an AUC of 0.94. However, the precision-recall analysis highlights areas for improvement, mainly in handling imbalanced datasets. This work demonstrates the potential of advanced neural models to drive better user engagement and increase sales in dynamic e-commerce environments while suggesting future directions for optimizing precision-recall trade-offs.