Automatic feature selection with ARO for brain signal classification

Brain-computer interface is a direct communication pathway between the brain and an external device. Electroencephalography (EEG) is the most studied potential non-invasive method to use in brain-computer interface applications, mainly due to its fine temporal resolution, ease of use, portability, and low set-up cost. The major challenges in brain-computer interface are feature extraction, feature selection, and increase recognition rate. In this paper, we used a new method for automatic feature selection which called asexual reproduction optimization (ARO). For classification, we apply three known classifiers, k-nearest neighbor, support vector machine and multilayer perceptron. Experimental results showed that with using ARO in feature selection we have 6 percent improvement in support vector machine and 13 percent for multilayer perceptron classifier. The best recognition rate which obtained is 93.43 for support vector machine classifier with using ARO feature selection

کلیدواژه ها:

Electroencephalography ، Asexual Reproduction ، Feature Selection ، Classification ، Support Vector Machine ، Asexual Reproduction Optimization ،