1- Islamic Azad University, Science & Research Branch, Tehran, Iran.
2- Full Professor Department of Medical Physics, Tarbiat Modares, Tehran, Iran.
3- Full Professor University of Tehran & ITRC, Tehran, Iran.
4- Associate Professor Department of Psychology, Tarbiat Modares, Tehran, Iran.
Abstract:
Electroencephalograph (EEG) signals reveal much of human brain states and this method can widely use in emotion classification. Although, the classification of emotion recognition is not almost ideal mainly due to the following reasons: (i) the features extracted from EEG signals may not solely reflect emotional patterns of a person and is affected by some time-varying factor and noise; and (ii) higher-level cognitive factor such as personality, mood, past experiences, etc. The dynamic properties of EEG data in relation to time series may affect the variability of feature distribution and interclass discrimination at different time stages. In this paper, we suggest a new adaptive ensemble classification method to alleviate the problems mentioned above. Specifically, we propose a new method for providing emotional stimuli. The Stimuli were sorted incrementally based on their valence- arousal score in three groups (sadness, neutral, and happiness).60 subjects 19–30 years of age (mean 25.01 and SD 3.13) participated in this study. The results show that the performance of emotion classifiers in this study has significantly improved compared to conventional classifiers. The classification accuracy elicited by the proposed method is 87.96 %.
Type of Study:
Original |
Subject:
Cognitive Neuroscience Received: 2021/11/26 | Accepted: 2022/04/12