Research on Detection Method of Unhealthy Message in Social Network

Abstract

In order to avoid the release and dissemination of eroticism, gamble, drug and politically sensitive message in social network, and purify the network space, we propose a method to detect unhealthy message in social network. Firstly, the Naive Bayes model is used to classify the message released by the social network. Then, according to the features of all kinds of unhealthy message, the classification model of Support Vector Machine (SVM) is used to make further judgment. The comparative experiment results show that, the classification model of SVM has better precognitive effect than that of Naive Bayes and Decision Tree.