Improving the efficiency of brain computer interfaces created on the basis of EEG signals

Al-Duhaidahawi MurtadhaAhmedLuti, Al-Nasrawi FarisHazim, Tomashevsky Yuri Boleslavovich

The paper is devoted to the topical topic of building a non-invasive brain-computer interface system (BCI). The way to increase the efficiency of BCI by solving the problem of mute speech, in which the subject speaks mentally without generating acoustic signals, is considered. EEG signals are used to recognize vowels. The process of training and classification of 5 groups of data using a periodogram, a decision tree and a support vector machine is described. An approximate view of the error matrix is given, and some recommendations for future work are given.

Keywords:brain-computer interface, electroencephalogram, support vector machine, error matrix.