Research of web-based user behavior analysis

• 1、Beijing Key Laboratory of Network System Architecture and Convergence, Beijing University of Posts and Telecommunications, Beijing 100876

Abstract：Network user behavior analysis has become a hot research topic. In the research of network user behavior analysis, network user identity which is an important aspect of network regulation, is of great significance in the management of network security. This paper analyzes the network user behavior to construct a model of the user identification, the goal is to achieve the identification of the network identity of the user. Network user identification analyzes the composition and the characteristics of the network users, researches the law of the network user behavior on the basis of the analysis of the network behavior, then identifies the identity of the user after extracting the relevant characteristics of the user behavior. The main content of this article is to study the methods of user identification. This article is based on the weighted naive Bayes classification (FWNBC) algorithm. The step is collecting relevant data, preprocessing the data, analyzing the characteristics of the network user behavior, extracting the feature which could reflect the user identification, using CHI feature selection algorithm to extract the feature, and then using the TF-IDF algorithm to weighted calculate the characteristic, and finally using FWNBC classification algorithm to identify the user. Experiments show that the model in the identification of the user identity is feasible and effective.

Keywords： network user behavior analysis user identification model FWNBC classification algorithm CHI algorithm