Three machine learning approaches in the epitope prediction

B-cell epitopes are important in both fundamental biological research and the clinical application. Here we get through three most widely used machine learning approaches (Naïve Bayesian Classifier, Support Vector Machine and the Artificial Neural Network) in the epitope prediction of both continuous and discontinuous types. As the prediction for conformational epitopes are still new in the epitope prediction, feature selection is especially analyzed. Some comparisons and the discussion about the advantages and disadvantages are made about the three methods.