The support vector machine (SVM) is a state-of-the-art method in supervised classification. In this paper the Cluster Support Vector Machine (CLSVM) methodology is proposed with the aim to increase the sparsity of the SVM classifier in the presence of categorical features, leading to a gain in interpretability. The CLSVM methodology clusters categories and builds the SVM classifier in the clustered feature space. Four strategies for building the CLSVM classifier are presented based on solving: the SVM formulation in the original feature space, a quadratically constrained quadratic programming formulation, and a mixed integer quadratic programming formulation as well as its continuous relaxation. The computational study illustrates the performance of the CLSVM classifier using two clusters. In the tested datasets our methodology achieves comparable accuracy to that of the SVM in the original feature space, with a dramatic increase in sparsity.
- Support vector machine
- Categorical features
- Classifier sparsity
- Quadratically constrained programming
- 0-1 programming