Study On Text Filtering Technology Based On RBF Neural Network

With the development of the computer technology and network technology, it provides more convenience for people to get access to the global information freely. At the same time, it also leads to information online grow explosively. Currently, users mainly use search engine to find information, it classifies information and creates database, then searches information according to the keywords, it returns the result to user and completes search. However search engine has its own location, the result exists useless information because of using keywords searching, so resulting in wasting time and space. So, I put forward a method of information filtering, using RBF neural network filtrates information that is searched according to key word, so that the result accord with the request of user much more.Filtering is one method to help users to obtain the information that mostly fit there needs. The function of information filtering is to select relevant information or eliminate irrelevant information from dynamic information flow on the internet according to certain criteria and some approaches. The thesis will expatiate the method of information filtering by using RBF neural network. First of all, I will deal with searched documents and extract eigenvector, then filtrating by RBF neural network, I will use self-organization algorithm and LMS algorithm to train RBF neural network. Then, the thesis proposes a design scheme of system model and describes the details of its implementation. Finally, I get the test result, it has very good information filtering performance.