Research On Intelligent Search Engine

As a main kind of method to retrieve information on Internet, search engine has been used in many fields. But the traditional search engine cannot meet people's demands on intelligent and personalized information service. So in this thesis, intelligent search engine is discussed, which can adjust itself to the user's interests and provides personalized information retrieval service.As a new solution, the information retrieval technique by natural language understanding and machine learning is presented in the thesis. The following aspects are investigated and discussed.A framework of intelligent search engine is presented. Chinese information processing model is added to the traditional search engine, which can make search engine intelligent and personalized.Chinese automatic word segmentation is the first work in Chinese information processing. In this paper, a Chinese word segmentation system is studied, which fits for intelligence search engine. Three principles and a solution for ambiguous word segmentation are presented.The application of artificial neural network to solve the problem of Chinese automatic word segmentation is presented. The mapping model and its performance are studied. Based on a number of experiments, the performance of the model is evaluated.Chinese document arrange and filtration is an important part in Chinese intelligent search system. This thesis introduces the method to filtrate Chinese document based on frequency statistics, the method developed from vector space model can get the correlative degree according to the user's interests and the effect is good.