Combination Methodologies of Text Classifier: Design and Implementation

Abstract

Support vector machines, one of the most population techniques for classification, have been widely used in many application areas. The kernel parameters setting for SVM in a training process impacts on the classification accuracy. Feature selection is another factor that impacts classification accuracy .The objective of this work is to reduce the dimension of feature vectors, optimizing the parameters to improve the SVM classification accuracy and speed. We present rough set method for feature reduce and a genetic algorithm approach for feature selection and parameters optimization to solve this kind of problem. We tried Reuters 21578 using the proposed method. Experimental results indicate, compared with the traditional methods, our proposed method significantly improves the classification accuracy and has fewer input features for support vector machines.