A Dynamically Adding Information Recommendation System based on Deep Neural Networks

Abstract

The recommendation system is a widely researched business tool that provides recommendations such as products or technologies that users are interested. The recommendation system learns user information by analyzing customer behavior and recommends products to meet customer needs. The existing mainstream recommendation systems still have room for improvement in terms of dynamically adding new users or new products to the system. The proposed method dynamically adds user and product information to the original recommendation system while recommending target content to new users or new products. While retaining the CIN and DNN structures, the networks associations are added to the before and after sequences. The input part of the add sequence is added to the previous sequence, the purpose is to dynamically update user and product information, adjust networks parameters based on sequence associations and learn high-order and low-order features of information. The results of comparative experiments show that our method could add new information dynamically with the low computing cost.