Research Of Query Expansion Based On User Behavior And Ontology

Query expansion technology can effectively improve the query efficiency in information retrieval. So Through taking the advantages of query expansion technology can greatly improve the search engine's efficiency of search accuracy. Query expansion technology can describe the user's information needs by adding the words related to the user to the user's original query. It can remove the query ambiguity and precisely search the user's required information.In order to improve the Query Expansion and overcome shortcomings of the traditional Query Expansion technology, such as bad semantic and bad topicality, this dissertation advance a new Query Expansion method based on the user behavior, combined with Ontology and Formal Concept Analysis. Three advantages can be integrated to form a set of query expansion technology based on formal concept analysis,user behavior and ontology. It can greatly improve SE search accuracy.The main research works of the dissertation are summarized as following:1. Advance a novel method of extracting the query expansion terms based on user's search behavior. A query expansion mechanism based on user's search behavior is proposed. The method analyzes their potential relevance of the user's search history and click history. The search return sequence and the user clicks sequence are considered as well. So we can extract terms which can depict the user's requirements. A concept lattice is built by using the web extracted terms. We consider the lattice the query expansion source. An experiment illustrates that our method can effectively extract the relevant terms and improves the quality.2. Propose a method of building user ontology. Extracting semantic relations (equal, similar, ISA and part of) of query and query source. Building user ontology with query, query source and relations using the method of ontology building.3. Propose a novel Query Expansion method based on the Domain Ontology. By using the semantic of ontology, such as synonymous, equal, ISA and part of, we present a new method to calculate the similarity between the new query and concepts in Formal Concept Lattice based on ontology. A new concept, which is most related to the user, can be found to expand the query by the quantitative method. So it can improve the search accuracy. Comparing the query expansion results of my method and the other ordinal methods, the experiment indicates my method is feasible.