Chinese Spam Data Filter Model in Mobile Internet

Abstract

With the rapid development of mobile Internet and the popularization of mobile intellectual terminals, learning, working and living were getting more and more efficient. While people were enjoying the efficiency if these modern products, large amount of spam information like advertising, pornography, gambling, fraud were flocking into people’s daily life. The information mentioned above was mostly transmitted through networking data targeting any application. Different methods of detecting and filtering had been proposed by many researchers and scholars, among which the method of SVM based on content was the most popular. However, research focusing on network data had rarely been conducted. The paper proposed a networking data filter running on mobile terminals based on Bayesian Classification. It focused on the data of Chinese spam data detecting and filtering. In addition to the automatic learning of new rules, the filter also implemented a passive incremental learning. Experiments showed that the filter had a higher detection accuracy and an average level of system resource occupancy.