Research And Realization Of Full-Text Search Technology

With the rapid increase of network information resources, there appears more and more professional websites. It has become more and more important that how to take comprehensive and accurate information from numerous network information within these websites to facilitate problem-solving of users. Search engine technology solved the difficulty of users searching information from the network, and full-text search technology is becoming the target that computer science and information industries are competing to research and develop.In order to meet the need of the actual requirements of the technology transfer center website in Beijing University of Technology, the thesis gave an in-depth and systematic research on the application of the technology transfer center website, and provided multi-faceted and more accurate information for the users through full-text search system.The thesis first researched full-text search technology in details, gave an in-depth research on the technology and the basic principle of full-text search, labored the structure of full-text search system as well as the framework of index, database structure and the creation process, and put forward an optimized method of index creation, greatly quicken the speed of reading of the temporary documents and increased the creation speed of index by mapping the temporary documents to the virtual memory. In addition, the thesis stressed on research and conclusion about the four models of search, sort algorithm and Chinese segmentation technology, and improved the maximum matching algorithm, fully realized the"long-term priority"principle against the lack of dictionary segmentation. The thesis also carried out the detailed analysis of the common used full-text search saddlebag Lucene, and compared with other open-source full-text search methods.The thesis also analyzed and researched the typical MVC model on the J2EE platform and its concrete realization - Struts framework, the principles of MVC framework, the basic components and the operating mechanism of Struts framework.Finally, the thesis discussed the design target of the full-text search function in the technology transfer center website, and designed the structure of full-text search system and all function modules. And the function modules include module design of static and dynamic pages, optimization of segmentation technology, improvement of Lucene sorting algorithm, as well as the design of dictionary in the segmentation engine and index of the website. Through the method of optimization of segmentation, it combined single Chinese word segmentation technology and dictionary segmentation technology, which brought about the advantages of good correlativity and high rate of search. It increased the score of links to websites and important information in the website through the improvement of Lucene sort algorithm, and the accuracy of the search system in the website. The thesis realized specific functions according to the final overall design and the design of various modules, and then it deployed the test operating of the actual website.