| With the increasing of Internet information,the new requirement for people wants to get more correct,detailed and deep information.So the Vertical Search Engines emerge as the times require.The paper takes the users' Auto Information search demand as the researching background.The Auto Information searched by Meta Search Engines is as collection object with the vertical search technology.The paper modifies the Nutch system framework based on full-text search software package named Lucene,designs and implemented Auto Information Vertical Search Engines which has Professional search function.The main work is as follows.(1) Reviewed the history of the development of search engines and categories; analyzed the problems faced General search engines;reflected Vertical Search Engines own unique superiority when it is applied in professional search area;gave the study status of Vertical Search Engines,and introduced the Lucene toolkit used in the system.(2) Analyzed the key technologies of Vertical Search Engines,such as Spider and Web Information Extraction,and focused on Correlation algorithm to determine the theme,Web Search Strategies and HTMLPaser Information Extraction.(3) Studied Lucene toolkit,focused on its index and Query Module;introduced calculation system of sort;and introduced the relationship between Lucene and Nutch.(4) Gave the design of vertical search engine on the automotive information, including system design,technology strategy,system architecture framework, development environment and so on;gave the system dynamic model and use cases.(5) Implemented three modules such as information capture of theme,information analysis and index and information retrieval of the theme.Tested the system and showed the results. (6) Gave a smart recommendation algorithm of search term based on continuous Hopfield neural network.Practice shows that,the design and realization of the vertical search engine for Auto Information in this paper is feasible,and the system had met the basic design goal. |