Design of an Intelligent Recognition and Filtering System for Spam Based on Multinomial Naive Bayes

With the rapid development of the times, the proliferation of spam SMS seriously affects the normal use of cell phone users and the safety of personal property. Therefore, in order to provide a cleaner usage environment, it is necessary to identify the characteristics of spam SMS and filter them effectively. In this paper, we compared the methods of Naive Bayes, Multinomial Naive Bayes, Bernoulli Naive Bayes and Gaussian Naive Bayes, found that the Multinomial Naive Bayes algorithm has the characteristics of fast speed, simple algorithm and high classification accuracy, so we chose the Multinomial Naive Bayes classifier as the core of the spam SMS filtering system. The intelligent recognition and filtering system was designed and implemented using Python programming language. The experimental results show that the system achieves between 84.00% and 98.00% in terms of the overall accuracy of SMS classification as well as the recognition rate of normal SMS, which helps the user to classify and filter SMS effectively.