Meta-search Engine Key Technologies And System Implementation

With the explosive increase of network inlbrmalion, it is more difficult for peple to look up information. The occurrence of the Web search engine overcomes this problem in some degree. However because different search engines use different scope and algorithms, the repetition of search results for the same query is no more than 34%. If wish to get relatively fullscale, accurate search results, multi-search engines should be used and the Meta search engines occur.How to make the Metasearch be used in a human way is the development of future Metasearch. We think how to select member search engine and how to have the search result auto-Cluster is critical technique of making Metasearch used humanly.In relation to how to select member search engine, we compared existing method. We estimated the performance of each method. In the end, we put forward some questions in the selecting of member search engine to be solved.In relation to the clustering of searching results, we first generally introduced the background of document clustering. Then, We put forward a method of n-Gram based auto-clustering that is fit for real-time implementation and multi-language.In the end, we designed and implement a parallel Meta search engine system. In relation to other Metasearch, this system can be used in a human way. and we analyze the query time and the precision ration. By compare to common search engine, we proved that this system is better than common search engine in precision ration.