Genetic Programming-Based Feature Selection for Emotion Classification Using EEG Signal

The COVID-19 has resulted in one of the world's most significant worldwide lock-downs, affecting human mental health. Therefore, emotion recognition is becoming one of the essential research areas among various world researchers. Treatment that is efficacious and diagnosed early for negative emotions is the only way to save people from mental health problems. Genetic programming, a very important research area of artificial intelligence, proves its potential in almost every field. Therefore, in this study, a genetic program-based feature selection (FSGP) technique is proposed. A fourteen-channel EEG device gives 70 features for the input brain signal; with the help of GP, all the irrelevant and redundant features are separated, and 32 relevant features are selected. The proposed model achieves a classification accuracy of 85% that outmatches other prior works.

Automated summary

The COVID-19 pandemic and resulting lockdowns have had a significant impact on global mental health. Therefore, recognizing emotions has become a crucial area of research for scholars worldwide. In this study, a feature selection technique called Genetic Program-Based Feature Selection (FSGP) is proposed. Using a fourteen-channel EEG device, 70 brain signal features are inputted, and with the help of Genetic Programming (GP), irrelevant and redundant features are separated, leaving 32 relevant features. The proposed model achieves a classification accuracy of 85%, outperforming previous works.