An enhanced support vector machines model for classification and rule generation

Abstract

Based on statistical learning theory, support vector machines (SVM) model is an emerging machine learning technique solving classification problems with small sampling, non-linearity and high dimension. Data preprocessing, parameter selection, and rule generation influence performance of SVM models a lot. Thus, the main purpose of this chapter is to propose an enhanced support vector machines (ESVM) model which can integrate the abilities of data preprocessing, parameter selection and rule generation into a SVM model; and apply the ESVM model to solve real world problems. The structure of this chapter is organized as follows. Section 11.1 presents the purpose of classification and the basic concept of SVM models. Sections 11.2 and 11.3 introduce data preprocessing techniques, metaheuristics for selecting SVM models. Rule extraction of SVM models is addressed in Section 11.4. An enhanced SVM scheme and numerical results are illustrated in Section 11.5 and 11.6. Conclusions are made in Section 11.7. © 2011 Springer-Verlag Berlin Heidelberg.