Combining Multiple Support Vector Machines using Fuzzy Integral for Classification

Recently, in the area of pattern recognition, the concept of combining multiple support vector machines (SVMs) has been proposed as a new direction to improve classification performance. However, current commonly used SVMs aggregation strategies do not evaluate the importance of degree of the output of individual component SVM classifier to the final decision. A method for multiple SVMs combination using fuzzy integral is proposed to resolve this problem. Fuzzy integral combines objective evidence, in the form of a SVM probabilistic output, with subjective evaluation of the importance of that component SVM with respect to the final decision. The experimental results confirm the superiority of the presented method to the traditional majority voting technique