An ensemble approach for brain computer interface applications

In recent past, Brain Computer Interface (BCI) has emerged as one of the fastest growing technology in the field of science and technology. With the continuous and dedicated efforts by many researchers, application of BCI technology has not only proved significant for disabled but also for healthy individuals. Here we have discussed about one such well known BCI paradigm i.e. P300 speller. The conventional P300 speller is actually based on detecting P300 signal through EEG recordings, which is supposed to occur when the subject sees targeted character or task. In this paper, we have discussed about the ensemble machine learning approach to detect the P300. Our main objective is to discuss the significance of ensemble techniques to classify and predict Event Related Potential (ERP). We have used two such well known ensemble techniques; Random Forest (RF) and Ensemble Support Vector Machines (ESVM). Finally we obtained ceiling level classification accuracy with minimum number of trials. These proposed techniques attest the viability for possible BCI applications using simple ERPs.