Application Of Test Classifier Based On SVM In Public Security Information System

In recent years, the rapid development of Internet technology makes the use of computers become more and more popular, people get more knowledge and information from the Internet commodiously. The mass of data gives the public convenience at the same time, but also gives the public security system of police officers a heavy workload. As we all know, in order to avoid the further spread of negative information, the police must deal with the illegal network information effectively. But there are many problems in the traditional manual processing method,such as manpower shortage, processing not in place and so on.In order to solve the problem of shortage of police resources 、 heavy workload and low efficiency, this paper designs and implements network information classifier based on SVM in public security system. Experimental results show that the classifier has achieved satisfactory results.The main research contents and achievements of this paper are as follows:(1) The topic in the phase of text segmentation, word segmentation program uses hidden Markov model(HMM) and use the python “numpy” library. This method can effectively reduce the loss of information in the process of Chinese word segmentation, so as to improve the efficiency and accuracy of the segmentation.(2) In the process of feature selection, the final choice of the feature extraction is realized using root test, after comparing several feature selection algorithms. Root test uses the difference measure formula to determine the degree of deviation theory value and the observed value, greatly reduces inexact generated error measurement, so as to improve the classification accuracy.(3) In the process of feature weight calculation, according to the special requirements of the public security system to the network information processing, the introduction of the intersection coefficient, improves the classical text feature weight calculation formula(TF-IDF).