ABSTRACT
Emotion detection is a booming research field in brain–computer interfacing, where researchers are trying to find various mental statuses of people in different situations, such as while watching movies, listening music, etc. The interfaces like mouse and keyboards are used to communicate with computers via human hands, similarly an emerging field called “brain–computer interfacing,” where the human brain can be used to communicate with the computer. Various signals like electroencephalogram, electrocardiogram, etc. are captured for detecting different categories of emotions. Among them, electroencephalography is the most widely used signal as it leads to suitability of different machine learning tasks. This work consists of (i) emotion classification using some suitable classification algorithms (support vector machine, K-nearest neighbors, and random forest) and (ii) transitions from one emotional state to another after a fixed time interval using the finite state transition machine. In this chapter, we have classified four classes of emotions (positive, negative, depressed, and harmony) based on electroencephalography signals input collected from brain activities. The proposed state transition system is associated with an increased or decreased value of related channels/electrodes of the specific cortex (frontal, temporal, occipital, etc.). After a performance analysis of the proposed model, we have also compared our proposed model with some state-of-the-art emotion classification models to measure the efficiency and effectiveness of our proposed system, and finally, the result shows that our proposed model substantially outperforms all previous models.
