Research on SVM and apply to network information content audit

Key technologies of network information content audit are researched,a model of network information content audit system are proposed,and a detailed description about the process of how to apply SVM algorithm to network information content recognition is given including feature selection,weight adjustment and so on.On this basis,SVM-based network information content audit system is developed.The test result indicates that the system can correctly identify sensitive content,recall and precision rates reached 95.06% and 85.47%,which solved the problems of selecting key words by manual way and meet the monitoring requirements.