Internet User Behavior Analysis Based On Web Log

It is becoming more popularity with the development of Internet technology rapidly,and the fiercer competition among Enterprises. It can acquire an advantage position inthe marketing competition who can understand the user’s browsing, using behavior andthe browse path deeply. And internet companies pay more and more attention to how torefine market positioning and provide more personalized services to the users. It canprovide guidances and advices for the enterprise to accurately marketing, and improve their marketcompetition, through exploring the behavior mode behind the data of the Internet users’ log. AndMining interest features of the different user groups can recommend more suitablefriends to communicate for the same user group, and provide a basis for widening theirsocial circle.The paper describes the concepts, processes and tasks of Data Mining, the concepts,processes, classification, commercial applications of Web Mining, K-means clusteringalgorithm, Prediction strength, the situation of China’s Internet development, theclassification and characteristics of Internet user behavior, the Data Mining application inthe analysis of user behavior. In this paper, I use the Statistics and Data Miningtechnology to analyze and excavate the users’ behavior, and find the pattern and mode ofusers’ behavior hidden in the data. The empirical parts of Internet user behavior based onweb log contain four modules, data distribution, data preprocessing, model building,interpretation and analysis. Related algorithm is achieved by the R statistical Software.In the paper, I classify and analyze the website and clients based on web log, andintroduce the prediction strength as the selection criteria of the values of k in order toestablish the model. According to the model, the paper summary the user habits andinterests characteristic of different user groups, and look for the interesting mode toprovide suggest for the enterprise to do personalized marketing activities.