E-commerce Recommendation System Based on Multi-Criteria Decision Making and Deep Neural Networks
Recommendation systems play a crucial role in e-commerce portals.Existing recommendation algorithms typ-ically rank products based on a single rating criterion,neglecting the comprehensive modeling of user and product char-acteristics from different criteria of user preferences.Therefore,an e-commerce product recommendation system based on multi-criteria decision(MCD)and deep neural networks(DNN)is proposed.First,a context-aware DNN model is designed to predict ratings from different criteria and obtain an aggregated rating prediction through an aggregation function.Subsequently,a hybrid model combining residual convolutional neural networks and bidirectional long short-term memory(Bi-LSTM)is employed to predict user sentiment towards products based on user reviews.Finally,the rating predictions and sentiment tendencies are integrated to achieve accurate product recommendations.Experimental results indicate that the proposed method achieves a mean absolute error(MAE)of 0.953 and a root mean square error(RMSE)of 1.129 onthe Amazone-commerceproductdataset,outperforming other comparative methods.This dem-onstrates that integrating MCD and sentiment analysis in recommendation systems can effectively enhance their perform-ance.
E-commerceRecommendation systemMulti-criteria decisionDeep neural networkSentiment analy-sisBidirectional long short-term memory
NETL
NSTL
万方数据
