Study Of Search Method Based On Group Characteristics

The Popularization and application of Internet is very rapidly, it is becoming an important way of obtaining information for people. The Internet architecture has the characters of open, heterogeneous and distributed, and the information in Internet has the characters of massive, redundant, update soon and so on. The search engine is one of the most effective way to obtain information, but with the development of Internet and user's information requirements, it is still has great difficult to obtain valuable information on the Internet. Generally, different users has different search intention and information requirements, and it has a certain specific and exclusive. General search engine technology does not take into account the different information needs, to the same search keywords, it return the same results, the search accuracy is poor, could not meet people's information needs. According to the user's interest to establish user interest model, introduces the user interest model into the search engine can achieve the goal of enhances the search accuracy. In view of traditional search engine system's limitation, the characteristic of people's information needs, this article carried on the search engine technology research based on the group characteristic user interest model.The main research work in this article has the following several aspects:(1) Analyse and study the general search engine's principle, structure and the composition, as well as the development present situation, it gives a thorough analysis to the existence question, has laid the foundation for the following research.(2) Construct user interest model based on the group characteristic, designed the search engine system frame based on the user interest model. Conduct the key research to the user interest model, and study the user interest model expression and the establishment technology. Based on the analysis and improvement of the fuzzy ISODATA algorithm etc.text clustering algorithm, it is gives the user interest model updating algorithm, the model can reflect the user group interest characteristic real-timely and accurately.(3) Study the search intention analysis and the expansion correlation technique based on the user interest model, then we propose the query extension algorithm, the results show that the efficiency of search engine is improved greatly.(4) The ranking algorithm of search engine has been studied based on the user interest model, the algorithm combine the user interest model to computation search key word and search result similarity, according to the similarity to rank the search results, enhances the sorting effect.(5) We design the prototype system of search engine and divide the functional module, based on group user interest model. The major function of the system has been realized by using the source software and tool such as Lucene and Java. All the functions of search system are emphatically researched and verified, the experimental results testifies the rationality of the design, the experimental results demonstrate the validity of this approach.Experimental results show that it can improve the performance for search engine, that the group user interest model is introduced into search engine, it testifies the rationality of the design, and the validity of this approach. But the method also has many defects, so it must be improved in the future.