|
With the rapid growth of the amount of information in Internet, the multitudeinformation has no organization, wide distribution, complex structure, dynamiccharacteristics; the search engine as the information retrieval tools has become anindispensable part of people’s life. The different of Independent search engine retrievalmechanism, method, scope resulting in the repetition rate as a retrieval request in differentsearch engine results in less than34%. Lawrence’s research shows that, search engineindex Web pages are not more than1/3of the total number of pages for any one.Therefore, to obtain comprehensive, accurate search results, more in line with theneeds of the user, it must be retrieved repeatedly calling different independent searchengine, but the weakness of independent search engine is born with a fact that cannot bechanged in terms of precision and recall.Meta search solve these problems in a certain extent, this makes the meta searchengine has been the academic attention and rapid development. Meta search engine is theintegration of multiple independent search engine results, has greatly improved in therecall, but also has exposed new problems repeated redundant information, noise, lowsystem precision, and meta search engine still did not consider the user personality factors,a search request to be applicable to all users, the search process does not take the userpersonality factors, the results of information retrieval can not meet the personalizedneeds of each user, which directly affects the quality of information retrieval and retrievalcost. In order to enable users to retrieve the most relevant information they need morequickly, and reduce the interference of useless multifarious information, the paper of targetis researching personalized meta search engine.This paper first introduces research status of the current development search engine;in view of the existing problems of numerous independent search engine, described thestructure framework with personalized meta search engine users personalized needs; In thepersonalized meta search engine and on the basis of in-depth analysis of research learningmethod based on user model, improved the search engine scheduling algorithm search, putforward the concept of member search engine results contribution and multidimensional sorting block, put forward the sorting algorithm based on user interest and search keywordsand search results related to the degree of the integration of text.Finally, this paper designed and implemented a personalized meta search enginesystem, and analyzed and test the proposed algorithm, compared to the general meta searchengine, the query is more humanized, compared with independent the recall and precisionof search engine has increased substantially, improving the user’s retrieval quality andsearch experience.
|