Research On Key Technologies Of XML Social Book Search

The traditional HTML-based information retrieval has been unable to cope with thecomplex information requests for users due to the diversity of data information onnetwork. For its scalability, simplicity, openness, interoperability, and many otherfeatures, XML is replacing HTML gradually and widely used in many fields.Especially in the digital book retrieval. XML-based information retrieval has become oneof the hot issues of international research in the field of information retrieval.Book is the most important way for human to access for knowledge. With thenetwork information and book digitization, people tend to get book materials on the web.However, using the traditional book search technology is difficult to get the book that theusers want to retrieval quickly and accurately form the large amount of information onthe network. In addition, the traditional book search system is based on the professionalmetadata to retrieve mostly, which requires users to have professional skills. Therefore, itcan’t meet the needs of the user with different cultural level and can’t discern intentqueries of different users at different times.In this paper, the key technologies of XML social book search are studied. We put thesocial metadata into book metadata indexing, take advantage of the user’s socialrelationships in query expansion and book recommendation. First, build inverted indexfor XML metadata documents of books. Secondly, proposed the method of queryexpansion based on user’s social relationships, divided the user’s social relationships intothree types: user’s interested objects, user’s friend objects, combine the user’s interestedobjects with the user’s friend objects effectively. And experiments were carried out onthe real book data set. The experiment results show that the query expansion method canimprove user satisfaction and the system precision effectively. Finally, analyzed the shortof traditional user-based collaborative filtering method and improved the method ofsocialization book recommendation based on collaborative filtering.