Contrastación de algoritmos de aprendizaje automático para la clasificación de señales EEG.

The purpose of this article is to present the parameter adjustment and the performance contrasting of two machine learning algorithms: support vector machines and artificial neural networks. Both algorithms are applied to the classification of EEG signals with motor imagination to discriminate the intention to open and close the hand. For this study, an own database was obtained through the Emotiv EPOC + device with fourteen signals. By means of a factorial experiment design and a statistical analysis, the parameters for which both algorithms present a better performance were obtained. Likewise, the most suitable algorithm for the classification of EEG signals was determined according to their accuracy, positive precision, negative precision, and false positive and negative rates.