A novel differential evolution algorithm integrating opposition-based learning and adjacent two generations hybrid competition for parameter selection of SVM

Generalization performance of support vector machines (SVM) with Gaussian kernel is influenced by its model parameters, both the error penalty parameter and the Gaussian kernel parameter. The differential evolution (DE) algorithms have strong search ability and easy to implement. But it falls into local optimum easily. Hence a novel differential evolution algorithm which integrating opposition-based learning and hybrid competition between adjacent two generations is put forward for parameter selection of SVM (DGODE-SVM). In DGODE-SVM algorithm, opposition-based learning and hybrid competition between adjacent two generations are inserted into the differential evolution process. Nineteen experimental results on UCI datasets distinctly show that, compared with ODE-SVM, SaDE-SVM, DE-SVM, SVM, C4.5, KNN and NB algorithms, the proposed algorithm has higher classification accuracy.

Automated summary

The proposed DGODE-SVM algorithm improves parameter selection for SVM by integrating opposition-based learning and hybrid competition between adjacent two generations. Experimental results demonstrate that it outperforms other algorithms in terms of classification accuracy.