Original scientific paper Electroencephalogram (EEG) signals have been considered to diagnose several brain and neurologic disorders. Moreover, the brain generated characteristic EEG signals according to the situation. Therefore, EEG signals have been used to detect emotional state and several EEG-based automated emotion detection models have been presented in the literature. In this work, a new automated EEG emotion detection model presented using multilevel discrete wavelet transform, local binary pattern, neighborhood component analysis, and k nearest neighbor classifier. The phases of the presented EEG classification model are; (i) the used EEG signals are divided into five equalnon-overlapping segments, (ii) frequency coefficients are generated using multilevel discrete wavelet transform, (iii) local binary pattern generates features from raw EEG segment and frequency coefficients, (iv) feature selection using neighborhood component analysis, (v) classification and (vi) hard majority voting. We used the GAMEEMO dataset to test our proposal. This EEG emotion corpus contains 14 channels and channel-wise results were calculated. Our proposal reached perfect classification rate (100.0%) on the GAMEEMO dataset. Moreover, the average accuracy value obtained from all channels was obtained as 99.36%. These results clearly denoted the high classification ability of our model on the EEG signals for emotion classification.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.