Research Of SNS Website Service Model Based On Web Mining

SNS refers to Social Network Site. In recent years, the Internet has been overwhelmed by SNSs. SNSs in China pose a relatively late emergence compared with those abroad and the development of them is based on the development model of Facebook. Therefore, there are many problems during the process of the development of China’s SNSs. For instance, they are experiencing problems of homogenization, the loss of users, the singleness of product functions etc. This thesis will conduct analysis by taking an early created and iconic SNS in China, renren.com, as example.Web Structure Mining is an important aspect of the technology of Web Mining. It is a technology applied to link analysis and plays a decisive role in the field of data mining. Frequently-used algorithms in link analysis include PageRank and HITS. By comparing these two algorithms, the thesis chooses the PageRank to be its algorithmic foundation of the research, and by analyzing the insufficiencies of PageRank, the thesistends to apply a modified algorithm, In-PageRank, to the research. In the modified algorithm, the weight between links is also taken into consideration rather than simply divide it equally. By comparing the previous two algorithms through experiments, the thesis testifies that the modified In-PageRank algorithm is more persuasive and precise than the traditional PageRank algorithm.As for homogenization and other problems in SNSs, the thesis tries to construct an SNS Site Service Model to solve those problems. First, In-PageRank algorithm is applied to the calculation of the PR ratings in main pages of renren.com, and the PR ratings are categorized and sequenced according to their value. Then the services provided by pages with higher PR ratings will be selected as the First Class services, and services provided by pages with lower PR ratings as the Second Class services. The service model is built with the idea of classification service, and in the meantime, personalized information platforms are also added to the service model. By providing users with personalized service, users’stickiness to the sites will be enhanced and their activeness improved.