The Research On Deep Web Database Based On Semantic Similarity Calculation

With the rapid development of the Internet, we are in an era that the amount information, which often contains a huge potential of knowledge. Although many engines have found a partial solution to solve the problem of resources, it is far away from satisfactory because of the poor degree of the relation and accuracy of the information. Some search engines, such as Baidu, Google, Yahoo, etc., often use the crawler technology to craw the website which related to the keywords to their own servers.When we search the keywords, we search the local files, we get some static pages of data.It is difficult for us to obtain the data hidden behind the Deep Web database, as a result, we lose a vast of valuable data.This paper presents a method of deep web database query based on semantic similarity calculation, the method aims to combine the semantic similarity calculation with the traditional search engines. Comparing with some method of calculating the semantic similarity, the essay presents an improved semantic similarity calculation which based on Tongyici and HowNet. By calculating the semantic similarity between keywords and the columns of the Attribute Dictionary, On the basis of a threshold-based matching algorithm. It determines the scope of the current keyword search query, and digs out the valuable information about the keywords in Deep Web Database. In the end, it generates the SQL, then show the datasct to the user.Sincu there is little test set of Chinese Deep Web DB. we use a DB of management information system of a university. The part of experiment takes examples to validate and test the feasibility and accuracy of our presented method. Our method not only avoids the cost of a full table scan, but also accurate the scope of the current keyword query. In the end, the combination of the method and the search engines, will improve the query experience.