Elephant herding optimization algorithm for support vector machine parameters tuning
Eva Tuba, Zorica Stanimirović · 2017
Classification is part of various applications and it is an important problem that represents active research topic. Support vector machine is one of the widely used and very powerful classifier. The accuracy of support vector machine highly depends on learning parameters. Optimal parameters can be efficiently determined by using swarm intelligence algorithms. In this paper, we proposed recent elephant herding optimization algorithm for support vector machine parameter tuning. The proposed approach is tested on standard datasets and it was compared to other approaches from literature. The results of computational experiments show that our proposed algorithm outperformed genetic algorithms and grid search considering accuracy of classification.