A Hybrid Recommendation Approach Integrating Matrix Decomposition and Deep Neural Networks for Enhanced Accuracy and Generalization

This paper proposes a hybrid recommendation system optimization method based on matrix decomposition and deep neural networks. By combining the advantages of matrix decomposition technology and deep neural networks, the proposed model can better capture the potential relationship between users and products and effectively improve the accuracy and coverage of recommendations. Experimental results show that compared with traditional collaborative filtering, singular value decomposition, and single deep neural network models, the proposed model performs better in terms of accuracy, recall, and root mean square error. This method provides an efficient and accurate solution for product recommendation tasks, with strong generalization ability and application prospects.