Query Conceptual Graph Expansion Based On Log Mining

With the development of the Internet and the information technology, the search engine is more and more important as a tool for finding information. Nowadays, the core of the search engine is the Boolean model, which depends solely on the keywords that the user gives. It just tests whether the words match or not, not the semantic meaning of the words. Therefore the search engine cannot always give the search results that the user really wants. The traditional query expansion can't improve the performance of search engine substantially mainly because it just makes sure the key words is match, without considering the relations between the key words and the semantic meaning of the key words, so it can't express the user's query intension exactly and expend the user's query, it also can't take away the difference between the user's query intension and the search results fundamentally. As a result, it can't improve the precision and recall of the search engine.Because users cannot always express what they want to search exactly, so this paper introduces the conceptual graph into query expansion, and gives the implementation of the query conceptual graph expansion based on log mining system. This system expands user's query by conceptual graph, it gets the user's search model by mining the search log, then it expresses the user's query with conceptual graph, and introduces a method to match the query conceptual graph with the user's search model conceptual graph, at last, it will output the conceptual graph that math with user's query.In this paper we use conceptual graph to express the user's query, the advantage of conceptual graph is that it can express the user's query intension exactly; through this way it can give the query expansion that can make user feel more satisfied. It because the conceptual graph not only stores the key words, but also stores the relations between the key words and the semantic meanings of the key words. The method that using conceptual graph in query expansion considered the semantic meaning of the key words whether match or not, it also considered the relations between the key words, which make it can expand user's query in accordance with user's query intension. So this method can overcome the shortcomings of traditional query expansion approaches.