Based On The Campus Network Information Filtering System Design And Realization

A lot of bad information filled in the online world, and poses a serious challenge to the campus network management. For this reason, the campus network could provide convenience to users and also bring harm to human. Information filtering is a systematic approach, which could be able to automatic dynamic information on the network flow and template matching filter to filter out harmful information.Bayesian theory has been very good used in information filtering applications ,because it is better than KNN, support vector machines and other methods. But there are also ignored the links between feature words, ignoring the risks of false positives and does not support incremental learning mechanism and other defects. In this paper, trash and harmful information campus network problem, the introduction to the Bayesian information filtering and the problems for the proposed improvement program, and ultimately build a campus network based on improved Bayesian filtering system. This article includes:1. An in-depth research of the key technologies of network information filtering and the related filtering modelThe general model of information filters and classification algorithms are discussed at first and analyzing the problems existed in the current information filtering system. Then, we focus on the network data's acquisition and representation, the calculating method of the features weights as well as the matching and classification algorithms and the key match and feedback techniques.2. Introduced the information filtering with improved Naive Bayes to generate the filtering templateNaive Bayes algorithm for the traditional text information on classification and filtering, not adequately taken into account by the individual characteristics of filtering information, it is used with some limitations. Therefore, we consider the concept of minimizing costs, its combination with the Naive Bayes algorithm and filtering information based on the features to be made improvements, given an improved Naive Bayes text filtering algorithm.3. Designed and implied a network information filtering system based on Naive BayesWe will improve information filtering technology Bayesian algorithm for the actual test results, experimental data to prove the reliability and validity of the algorithm, achieved satisfactory test results.