A Probability Transition Matrix-Based Recommendation Algorithm for Bipartite Networks

摘要

By analyzing the searching and recommendation system in the electronic commerce field, this paper provides a retrieval sorting and recommendation algorithm for e-commerce service. An information transmission matrix is constructed based on the 'customer-product' bipartite networks, and the rankings of customers and products can be obtained by analyzing the network structures. Then we propose a community detection algorithm for bipartite networks, by employing the information transmission matrix. Finally, the recommendation scheme based on both customers and products is obtained. It makes recommendation results more comprehensive and reasonable, which satisfies the various requirements of the customers.