Tuning and evolving support vector machine models

Support vector machines (SVMs) are a well-established classifier, already applied in a variety of pattern recognition tasks. However, they suffer from several drawbacks—selecting their appropriate hyper-parameter values (the SVM model) along with the training sets being the most important. In this paper, we study the influence of applying various kernel functions in SVMs. We verify not only the classification performance of the classifier, but also the number of selected support vectors and the training time for each kernel. Also, we perform the qualitative analysis of the retrieved support vectors using an artificially generated dataset. Finally, we show how to optimize the SVM models using a genetic algorithm. An extensive experimental study revealed that evolved SVM models provide high-quality classification and are retrieved in much shorter time compared with the trial-and-error approaches.