A NB-based approach to anti-spam application: DLB Classification Model

Classification using Naive Bayesian (NB) classifier model, which is the context - based spam filter method, is a hot topic. The NB classifier is a simple and effective classifier, but its attribute independence assumption makes it unable to express its semantic relation. A new classification model is proposed that call Double level Bayesian classifier model (DLB). It not only considers the semantic dependence, but also has the simple and effective characters that are the advantages of NB classifier model. The conclusion we get from the experiment is that the performance using DLB classifier model is better than which using NB classifier model.