Existing Emotional State Detection Models: Review Paper

Our daily lives are filled with emotion as a subjective psychological and physical response to environmental events. Emotion recognition using EEG data is quickly emerging as a multidisciplinary study area with the advancement of artificial intelligence and brain science. This study covers the approaches for extracting emotional features from EEG signals and for classifying them using those features utilizing relevant scientific literature from the previous five years. Techniques for feature extraction analysis that are often utilized include time domain, frequency domain, and time-frequency domain studies. With classification accuracy ranging from 70.89% to 99.79%, algorithms for machine learning that use Support Vector Machines (SVM), k-Nearest Neighbour (KNN), Naive Bayes (NB), and other techniques are often utilized. Classification accuracy for the neural network (CNN), Random Forest (RFC), and XGBoost-based deep learning algorithms ranges from 65.38% to 80.56%.