Research And Design Of The Individuation Search Engine Based On The Web

With the explosive increase of information in web, it is difficult to search the needed information in information marine. So the search engine has become the main tool for information search. Although the search engine brings a great convenience for searching information, there are still many shortcomings in most of search engines. They do not consider individuality and interest to the consumer, the inquiry being able to only carry out consumer simple needs. As long as the keyword that the consumer uses is identical, what reconnaissance result identical has reduced the search the precision. Therefore, how the information resources collecting is organized rationally, in how secondary large amount of information, different specifically for the consumer interest needs, return to the information that the consumer needs really, realize the important problem individuation to search for, becoming to study at present thereby.The paper aims to the problem of search engine system, has gone deep into the systematic relevance of search engine technology studying realization individuation, has designed to carry out also individual search engine, and mainly be absorbed in individual the analytical organization searched the web page resources of engine, automatic classification and individual model of web page set up renewal etc. aspect come analytical research. The main work of paper is as follows:(1) Carries on the elaboration analysis to the present search engine system. The accuracy discussing to analyze to search the development history, system of engine system structure currently, and analyzing to searching some blemishes of system existence at present, search for example isn't high, can't body now the character of customer.(2)To individuation, Web page characteristic in search engine describes the calculative method having carried out the weight having studied, and bringing forward one kind of the word making use of the non-linear function to improve term weighting method.(3) The automatic classification of web page. The efficiency studied currently more popular classification calculate way, made use of classification calculate way to carry on a classification towards collecting a web page information a resources, contracted the search scope of customer from the certain degree, raised a search.(4)The individual model sets up. Adopt Web mining technique to save a medium history page to carry on excavation towards depositing in Web slowly, obtain the interest information of customer, make use of to gather a type of calculate way a classification a management to the customer interest, and make use of the form of two superior fork trees to mean interest in the customer. This article uses the gain the user interest information to construct the personalized model.(5)The Agent dynamic state follows with the renewal of individual model. Making use of the Agent dynamic state follows customer to browse behavior, catching the variety of interest in the customer, and pass "weak factor", establish the power heavy value, the interest worth and time in time worth to renew interest in the customer, renewing model continuously.(6) Making use of the individual model percolation searches a result, returning to satisfy and its interest demanding characteristic result for customer. Here put forward to result percolation calculate way.