Integration of Support Vector Machine with Naïve Bayesian Classifier for Spam Classification

In this research, we propose a two-stage method for spam classification, the naive Bayesian classifier (NBC) and support vector machine (SVM). NBC adopts the concept of Bayesian theory for classification, and combines the conditional probability with feature count as input data for SVM which uses the radial basis function with Gaussian kernel for further classification. The classification features generated from spam data set are used to train and test the proposed method. The results are compared with other well-known classification methods to verify the performance of our proposed classifier based on the precision and recall rate.