Research On Recommendation Technologies Of XML Faceted Search

Characterized by simplicity, scalability, opening and interoperability, XML has become the standard of Internet data representation and data exchange. With widely application of XML technology in all walks of life, the number of online XML documents shows explosive growth state. Effective information retrieval based on XML documents has become one of the hot topics of international information retrieval research. Traditional XML information retrieval belongs to the precise inquiry, which mainly depends on the keyword matching and Boolean queries, but most of the queries cannot be accurate to the specific queries; Facet search is an exploratory information retrieval technology, which is a complement for keyword search and makes up for the shortage of the traditional search method. Users can select a valid qualification, gradually narrowing the range in the result set until getting the desired information. Facet recommendation technology is the key of facet search system.In this paper, the recommendation technologies of XML facet search are studied. We improved the facet recommendation algorithm and result recommendation method.At first, the inverted index structure of XML documents were built, Secondly, the facet recommended methods of XML document were also researched in the paper, which analyze the facet selection information of historical users, mining the potential correlation between the facets by using the association rule, and then facet recommendation were made. The experiment results show that considering the relevance relationship between XML facets can effectively recommend the most navigational facet. Finally, the deficiencies of traditional method based on user collaborative filtering are analyzed, and then the book recommendation method based on collaborative filtering is improved, combined with the user’s social relations, which are applied to the search system of books.