Spam classification based on different artificial intelligence methods

Abstract

The increased need for social communication has led to an increase in email users, and with it more spam is being spread. In this paper, by comparing and exploring the accuracy of some supervised machine learning methods and a deep learning method called Long short-term memory (LSTM) on the problem of spam classification, this paper aims to provide more solutions for the problem of spam filtering. This paper firstly conducts an in-depth understanding and analysis of the principles of different machine learning algorithm models, which is very helpful for the following research. Then the experimental comparisons after mastering the principles of different models are conducted. Regarding the process of the research, the data set was first pre-processed to facilitate the use of different algorithm models, and then the data set was put into different models for training. Finally, by comparing the accuracy and confusion matrix, it was concluded that LSTM was used in spam classification. problem is more advantageous.