Research On Deep Web Dynamic Search

Based on features of book sail websites, the thesis designs he method that form items is parsed and filled dynamically according to the text before the input item reflect the information to be input in the input item, Make use of dynamically form to get result page, parse and sort results page by weight, at last display them according to the uniform display format. The paper designs and implements the system to query the same type of multiple websites based on their advanced search pages, at the same time the system provide convenient and efficient condition to query books on multiple book websites for users. Experimental results demonstrate the correctness of algorithm, the main research topics include:1. This paper has design a dynamic form search algorithm based on dictionary matching. The algorithm parses a form with SAX to avoid large quantities of useless information with existing DOM; improve processing performance with multiple threads to parse query interface page; make use of dictionaries to match key words of form items. On server side pages are crawled to make semantic analysis, to find new book sail websites and expand the book keywords dictionary. 2. Based on the results of dynamic filling the form, the paper realizes the result pages parsing. Through foreseeing the structure of the HTML tags on the search result pages, extract this tag structure with abstract extracted tags to get books information object linked list, and complete results analyses.3. The proceeding work of query results. To resolve the results of sorting in the result page, main consideration factors are the frequency the similar books in different websites appear and sorting in every website. Two factors are equally important, both reflect popularity and sales situation of books, so the paper uses equivalent weighted ranking.On the basis of above work, designed and implemented a library website search system based on advanced search pages. The system provides a relatively new idea, for the same type of website, precisely inquiry items through its advanced search page.