Research And Implementation Of Vertical Selection System Based On Semantic Expansion

As the development of the Internet, the information on the Internet is becoming more and more larger and ten of thousands of them are updating every day. To search the available information in the huge information base need a high-efficient retrieval tool. Furthermore network resource which is text formatting has become more diversification. Facing the multi-domain and diversified information data, vertical search which is also named field search has emerged. However the problem also emerges, if you know the information belongs to the fields, you can search it directly. But if you don’t know, it involved a problem that which vertical to choose when a query come. This paper just study on this problem.This paper adopts ontology semantic expansion, it will get the user’s query strings; and then the initial query will be expanded a set of concept which is around it through ontology semantic expansion. For the set of concept, we also design and implement inverted index which is established with query log, and use the inverted index to perform pruning. Thus the noise influence of query expansion becomes lower.Finally, for the vertical selection, there are two kinds of scores of document, Those are score of document of initial query and expanded. The score of initial query comes from the inverted rate and the word rate, and the other come from the click rate of the document and the sequence of the document. Through extracting these information from the log, we can get the weight of the document and its inverted rate, thus we can compute the score of the documents. In the end, we plus the two parts score as integrated score, and we use every integrated score as vertical score, and the integrated score is used for vertical selection.The experiment testifies that the system which contains ontology semantic expansion and inverted index is benefit for the accuracy of vertical selection.