Implicit Rating Model in M-Commerce Recommendation System
Hongwei Liu, Zhouyang Liang · 2009
Collaborative filtering technology is the key technology of recommendation system. However, collaborative filtering technology has been suffering from sparsity that it needs mass ratings from users to improve precision. In traditional e-commerce, asking users to rate on their own initiative will degrade experience of users, let alone the mobile business environment. So, both in e-commerce and m-commerce, it is very difficult to collect enough ratings. In this paper we will propose a novel model, Bayesian network-based implicit rating model, which intends to solve this problem. Browse behavior, marketing basket data, and context information will also be considered in a comprehensive way to construct a Bayesian network. In addition, the successful implementation of the model through experiment carried out in the mobile environment indicates us the plausibility of the model.