A Survey on SVM Hyper-Parameters Optimization Techniques

Abstract

Support vector machines can be considered as one of the most powerful classifiers. They are parameterized models build upon the support vectors extracted during the training phase. One of the crucial tasks in the modeling of SVM is to select optimal values for its hyper-parameters, because the effectiveness and efficiency of SVM depend upon these parameters. This task of tuning the values for the SVM hyper-parameters is called as the SVM model selection problem. Till now a lot of techniques have been proposed for optimizing values of hyper-parameters of SVM both in static and dynamic environment. Static environment is one where the knowledge about a problem does not change over time due to which static optimal values can be assigned to the hyper-parameters. On the other hand, due to the changing nature of the knowledge about a problem, in dynamic environment the optimization process has to be flexible enough to adapt the changes quickly. In dynamic environment, re-evaluation of the optimal values of the hyper-parameters is needed. This paper attempts to identify various optimization techniques used for SVM hyper-parameters tuning and recognize their pros and cons.