Extracting Patterns in Social Network by Means of Link Mining
Hnin Wut Yee Oo, Than Htike Aung · 2010
The World Wide Web serves as a huge, widelydistributed, global information service center for news,advertisements, consumer information, financialmanagement, education, government, e-commerce, andmany other information services. The Web contains arich and dynamic collection of hyperlink informationwhich can access usage information and provide richsources for data mining. It is a highly dynamicinformation source and web service centers update theirweb pages regularly. Linkage information and accessare also updated frequently. This framework is thedistillation of broad search topics, through the discoveryof “authoritative” information sources on topics. Thisusing hubs, called HITS (Hyperlink-Induced TopicSearch) to find authoritative pages based on therelationship between a set of relevant authoritativepages and the set of “hub pages” that join them togetherin the link structure. This formulation has connections tothe eigenvectors of certain matrices associated with thelink graph. This HITS algorithm provides surprisinglygood search results for a wide range of query.