Automatic Email Spam Classification Using Naïve Bayes

Abstract

The effectiveness of the Naive Bayesian machine learning method is examined for spam filtering. Reliable anti-spam filters are required due to the rising volume of unsolicited bulk emails (spam). Until now, most of the keyword patterns used in these types of filters have been manually created and have had poor performance. Recently, it has been suggested that using the Naive Bayesian classifier is a good way to build superior automatic spam filters. On a publicly accessible corpus, a dataset from Kaggle is used to investigate the Naive Bayesian filter's performance, contributing to industry benchmarks. Analyses of the Naive Bayesian filter's performance have also been done concurrently. This method outperforms a widely used e-mail Here, the effectiveness of the Naive Bayesian machine learning method is examined for spam filtering. Reliable anti-spam filters are required due to the rising volume of unsolicited bulk emails (spam). Until now, most of the keyword patterns used l reader's keyword-based filter in terms of accuracy and precision for the dataset under consideration (95.56% and 93.91%, respectively).