Research On Domain-Oriented Search Engines Based On User Behavior

As the continuous development and increasing popularity of Internet, the amount of information on computer networks is increasing at very fast speed. As a tool for people to search network information, Search engines are applied widely. But in the developing of network technique and applied need, the problems of new network information search appear continuously, search engines face new development trend. The users'need to the information looks toward to the development of diversification, no longer need information to some domain vast and wide, but accurate and deep; the users'request to the information isn't quantity any more, but the quality of the information, namely whether can provide accurate, authoritative information in time. So Personalized Search Engine and Vertical Search Engine which can satisfy people's particular need emerge as the times require.This text mostly aims at the problem of not strong speciality and individuation of current search engines, built domain model and user behavior model based on the deep analysis of users'action character and technology of search engines, and validate the feasibility of the idea of Domain-Oriented Search Engines based on users'behavior. This text discussed to the related arithmetic and key technique of the establishment of the domain model and the users' behavior model in the Domain-oriented Search Engines based on users'behavior. This text applied ID3 to build domain model, rebuild indexes based on topic with the foundation of the Index Database of General Search Engine, and use it to filter the indexes which have no relationship with the topic, and thus attain the purpose that it can retrench the Index Database and improve the searching efficiency. In the process of creating user behavior model, the behavior science theory is used for clustering and expressing users'search behaviors and generating the users'behavior model. Based on it, we reset the index of the Index Database; accordingly attain the purpose of accurate query. Finally, the Domain-Oriented Search Engines based on users'behavior is evaluated through the experiment. The experimental results show that the search results through Domain-Oriented Personalized Search Engine based on users'behavior are better in meeting the requirements of users.