| With the rapid development of Internet and electronic commerce,shopping becomes more and more convenient. On the one hand, different kinds of goods meet the manifold needs of people. On the other hand, the more goods there are, the harder for the consumers to choose. This leads to the truth that consumers find it tiring to do some shopping and even lost the interests on it. What mentioned above is called information overload. As one of the solutions to solve this problem, personalized recommendation system in electronic commerce fits every consumer needs and provides them with personalized products platform and shopping experience.In this paper personalized recommendation system in electronic commerce is mainly discussed. To begin with, both domestic and international research results are analyzed. Then, the popular recommendatory algorithm is also analyzed. Furthermore, a collaborative filtering algorithm based on latent factor model and clustering algorithm(CF-LFMC) is put forward. Finally, a personalized recommendation system on mobile terminals based on the algorithm is designed.The main work done in this paper has the following several aspects:1. The traditional problems on data sparsity, timeliness and cold start are analyzed which exist in the Item-based Collaborative Filtering (IBCF).The latent factor model and clustering algorithm are used in this paper,and the traditional IBCF is improved by introducing the time function and initial project attributes filling method. CF-LFMC is proposed based on the latent factor model and clustering algorithm.2. A personalized recommendation system is designed in electronic commerce. This paper describes the system architecture and designs from client to server and analyzes the inside modules in detail. At last, it introduces the realization process.3. CF-LFMC is verified by experiments. The results show that it is better than the traditional algorithm on accuracy rate, and it achieves the desired effect. |
PDF Full Text Request