Research of shopping recommendation system based on improved wide-depth network
Shanshan Wang · IOP Conference Series Materials Science and Engineering · 2020
Abstract With the development of e-commerce, there are more and more commodities. How to recommend the commodities that users are interested in quickly and accurately has become an important research topic in the field of e-commerce.we propose a product recommendation algorithm based on DeepFM network. Firstly, we embed the user’s purchased products, and transform the sparse feature into the low-dimensional dense feature, while the user’s personal attribute features can express the user’s purchase intention to a certain extent, and also use embedded coding to transform the features.DeepFM considers both wide and deep (i.e. low-level and high-level) at the same time to further improve the generalization ability of the model. So we use DeepFM to predict the interest of users in purchasing goods.Learning the expression of user’s interest from user’s purchase and personal preference, so as to accurately predict user’s purchase behavior.Finally, we use the real record data set purchased by online users to evaluate the effect of the model, and compare it with other models to verify the effectiveness of the model.