Research on Filtration System of Network Negative Information on the Basis of Naive Bayes

摘要

The difficulty of filtrating network negative information lies in how to classify information correctly. As one of the classification method with the advantage of strong robustness and good understandability in the field of pattern classification, Naive Bayes has been used widely. A method for filtrating network negative information on the basis of Naive Bayes, improvement proposals aiming at the disadvantages of Naive Bayes and amelioration of erroneous judgment of negative information by setting threshold value k have been put forward in this article. The experiment shows that by adjusting threshold value k can the integrity of the system can be optimum and can favorable application effects be achieved.