Eye State Classification Using EEG Based on Electrode Position

Initially research in the field of Brain Computer Interface (BCI) was to help people who have special needs, but over time BCI's utilization has become more widespread including education, robotics, prostheses, mechanical device controllers, and others. One of them is research on how to EEG to detect the condition of eyes with two-dimensional eyes closed or open. By using the public data set of the status of the eye, this paper aims to investigate whether the location of the electrode affects the quality of accuracy, precision, recall and area under the curve (AUC). To process data, the author uses a rapid miner device. The location of the electrodes observed were three areas, namely the left side of the brain (AF3, F7, F3, FC5, T7, P7, O1), the right side of the brain (AF4, F4, FC6, F8, T8, P8 and O2) and all parts of the brain (AF3, F7, F3, FC5, T7, P7, O1, O2, P8, T8, FC6, F4, F8, F4). The classification model, which is used, is logistic regression (SVM). Observations of data taken from electrodes in all parts of the brain are higher compared to data taken from electrodes on the left and right sides. For classification with the left side, the accuracy rate is 55.98%, the precision level is 64.03%, the level of 4.41% and AUC is 0.611%. For the classification with the right side, the accuracy rate is 55.13%, the precision is 50.67%, the level is 1.88%, and the AUC is 0.587%. For all parts of the brain is 58.71%, 79.14% precision, Recall is 10.91% and AUC is 0.618.