Research Of A Postprocessing Method Of Meta Search Results Based On Information Distribution

Human society has entered the information age. The increasing amount of information has brought the joy of sharing to us. As the increasing of the web information resource, a lot of useful information has been inundated so that finding the useful information becomes more and more difficult. In order to acquire the required information and make the internet service more personalized and intelligent, the search engine and personalized service should be provided.The advent of search engine also brings us a problem that for the same user but different time, most of the results are same. Meanwhile, in order to solve the requirements of different users, traditional method of setting up a personalized service mechanically applies the information from user's inquiry and the system to search, and the results have much duplication too. How to acquire useful information more conveniently and effectively?Basing on that requirement, the postprocessing method model of meta search results which was using meta search technology were put forward and accomplished.1. Setting up a user model according to the user's searching behavior and activity of the results.2. According to the results of this inquiry and historical results, information duplicate removal was put forward to exclude the results of high level similarity, provide the user non-repeated and updated information. It provides facilitated and agile renewed results for the user.3. In order to provide the user more suitable results, the Naive Bayes algorithm was put forward. 4. Using OWL to represent the data which was transferred in the modules and for the data saving. It's beneficial to the operation of the process and the research at the next step.The experiment which was performed to record the appearing time of the changed information showed that this solution provided preferable services to the user mentioned above compared with the traditional solution. Meanwhile, a comparing experiment between this model and common search engine was put forward in dealing with the distribution of information. Comparative results show that it is significantly improved by using this model, and meet the user needs.