Research On Web Data Mining Based On Social Network Analysis

With the rapid development of Internet, the number of pages on the web shows explosive growth. Although search engines can help us get pages relevant to topics in some degree, the search scopes of most search engines is small and a high percentage of the search results have nothing to do with the users’ requests. How to find the wanted pages accurately from more than a trillion pages has become an urgent problem. There is a kind of page called authoritative page which has high credibility and is linked by most pages relevant to topics. If the search engine gives authoritative pages when it gives search results, then it can greatly improve the search quality and search efficiency.Social network analysis is first used to analyze the relationship between human groups and search for core members of the groups. The link relationship on the web is similar to the human relations in the realistic society. This paper applies social network analysis to the web mining and studies social network analysis in application of web data mining.This paper first elaborates web data mining, then introduces social network methods in detail and compares all kinds of social network analysis softwares, eventually selects UCINET as the experimental analysis software. Meanwhile this paper introduces and develops the web spider used in the experiment.In this paper, we propose web data mining based on social network analysis and introduce the principles of the method. From the given keyword, we obtain seed URL from Google’s search engine which is based on the arithmetic of Page Rank. Then we use the web spider which is developed by myself to crawl URL as the given depth which are used in experiment.As well as we use social network analysis to mine core person, we apply Degree Centrality Analysis, Cohesive Subgroups Analysis to the URL we obtain, then mining the authoritative page which is given the keyword. The experimental results show that the web data mining based on social network analysis can mine the authoritative pages effectively.