Group Search Optimizer-based Neural Network for EEG-based Emotion Recognition

Abstract

Emotion recognition based on physical signals, particularly electroencephalogram signals, has developed into a prominent study area and gained a lot of consideration in recent years, with the continued progress of brain-computer interface and artificial intelligence technologies. Electroencephalogram (EEG) based emotion recognition has recently become critical in enhancing the intelligence of the Human-Computer Interaction (HCI) system. In this research, emotion recognition based on EEG data is used to develop and design the most accurate identification. The DEAP dataset is used as input signals to perform emotion recognition. The frequency bands for the EEG signals are generated as delta, theta, alpha, gamma, and beta, which depends on the activities as normal or abnormal consciousness with alternative frequencies. The artifacts are removed from the EEG signals in the initial stage, in the terminal stage, the frequency-based features are extracted from the generated frequency bands in the previous phase. The main intention of the developed Group Search Optimization-based Neural Network is to accurately predict the candidate emotions from the physiological-based EEG signals. The proposed method reveals better performance in terms of its accuracy, sensitivity, and specificity than the prevalent methods considered in this research.