Affiliations: [a] School of Mathmatics and Computer Science, Wuhan Polytechnic University, Wuhan, Hubei, China
| [b] School of Computer Science, Wuhan University, Wuhan, China
Correspondence:
[*]
Corresponding author. Xiaoyue Tang, School of Mathmatics and Computer Science, Wuhan Polytechnic University, Jinyinhu Machi Road, Dongxihu District, Wuhan, 430024, Hubei, China. E-mail: [email protected].
Abstract: The paper uses BP neural network method to analyze the behavior preference of web users to achieve user behavior clustering, help advertisers to find online advertising design and optimize web design. At the same time, the paper analyzes the user preference cluster analysis, which helps users to find the desired webpage and content more conveniently, shorten the retrieval time and improve the retrieval efficiency. Firstly, the paper analyzes the log files of the web server, and then conducts session classification, finds the frequent data from the session vector, and normalizes the generated pattern vector. The BP SOFM model is used to cluster the user behavior preferences to generate users. Clustering. The experimental results show that BP neural network analysis method can effectively analyze user preferences and cluster user behavior.
Keywords: BP neural network, cluster analysis, web user preference analysis, SOFM