Improvement and Implementation of Spam Filtering Based on Bayesian
Shuwei Zhang · Microcomputer Information · 2011
This paper analyzes the disadvantages of native Bayesian algorithm which has been widely used in anti-spam filtering.Moreover,the paper shows a new improved Bayesian filtering model in terms of pattern matching,fuzzy matching.The first preprocess of e-mails filtering is catching feature items——pattern matching which could distinguish junk emails from normal mails with a number of training e-mails.Furthermore,extract the feature items out of the e-mail with fuzzy matching.The experimental result demonstrates that the new algorithm applied has effectively improved the accuracy of spam filtering.