In this study, features extracted from EEG signals were analyzed using machine learning methods to facilitate the diagnosis of Major Depressive Disorder. 3-channel EEG signals in the MODMA dataset, obtained in the resting state of 26 Major Depressive Disorder and 29 healthy subjects, were analyzed. Each channel of these signals was divided into 6 different frequency bands. These bands were delta band (0.5-4 Hz), theta band (4–8 Hz), alpha band (8–13 Hz), beta band (13–30 Hz), gamma band (30–60 Hz) and alpha-beta band (8–30 Hz). Using the features extracted from each channel, classification was performed using machine learning algorithms such as Decision Trees, Support Vector Machines, k-Nearest Neighbour, Naive Bayes, Ensemble Classifier, and Wide Neural Network. At the end of the training phase, 95.83% accuracy, 95.32% F1-score, 95.83% recall, 94.47% specificity, 94.82% precision, 90.32% Matthews correlation coefficient, 90.32% kappa, and 0.0973 false positive rate were detected in the combination of alpha and beta bands (8–30 Hz) using the Ensemble Classifier model. It was determined that the combination of alpha and beta bands provided the highest accuracies.
