Application of Boosting Algorithm in Spam Filtration

Abstract

In order to improve accuracy and effectiveness of junk mail filtration, a filtering method is proposed based on Boosting algorithm, which uses Boosting algorithm to construct a spam filterer to identify junk mail. Besides, reference technical indexes in the field of text classification and information retrieval are used to construct a spam filterer evaluation system, with it, experimental data obtained from simulation were tested and evaluated. The results of test and evaluation proved that, compared with the traditional Bayesian algorithm, the spam filterer based on Boosting algorithm is able to filter junk mails more excellently, and the effectiveness of Boosting algorithm in spam filtering is verified.