Advancing Human Computer Interaction Through a Hybrid EEG-Based Emotion Recognition System

An emotional recognition system helps human-computer interaction systems by using a deep learning approach to find pattern recognition. This research advances the HCI field and lays the foundation for a new era of emotionally intelligent and responsive technology interfaces. The EEG emotion recognition system is consisting of machine learning algorithms for preprocessing, feature extraction, feature selection, and classification. The original raw signal contains noise and irrelevant signals to be removed using ICA (independent component analysis). The next step is feature extraction to extract features such as variance, standard deviation, kurtosis, and entropy from the EEG signal using the discrete wavelet transform (DWT). These features are selected based on the type of emotion using linear discriminant analysis (LDA). The hybrid classification algorithm uses CNN (convolutional neural network) and GRU (gated recurrent unit) to classify the emotion. The results of this hybrid classification model are precision, accuracy, and sensitivity for better outcomes.