改进的文本分类技术在垃圾短信过滤系统中的应用

With the development of short messaging service (SMS), the content and quantity of junk SMS are constantly changing. Some problems such as reduced precision rate and low speed rate exist in the traditional filter system of junk SMS. To address these two problems, this study focus on the application of Naive Bayes text classification in junk SMS. Considering that the processing rate of Na?ve Bayes text classification would be reduced under high dimension and mass data, I reduce the dimension through feature extraction and simhash. Meanwhile, to remedy the precision defect of Na?ve Bayes classifier, I use dual filter by combing simhash filter to raise the precision rate and also improve the effectiveness of filter system. The results show that the above solutions can effectively solve the two problems.