Analysis and Classification of Physiological Signals for Emotion Detection

Abstract

Detecting and classifying emotions using several physiological signals has become a pivot area of research nowadays. The most popular method for analysis of emotion recognition is the use of physiological sensors. This paper focuses on physiological signal-based emotion recognition, including analysis of emotional physiological datasets and classifier models. The study helps human computer interaction (HCI) research immensely. The acquisition of the signals through heterogeneous datasets is done through several physiological sensors like PPG, GSR, EEG, etc., to detect human emotions automatically by selecting best-fit algorithm. The signals in terms of training datasets are extracted once the analysis of the pre-processed data is over and is validated using data validation model. The trained and test datasets are classified based on some machine learning models that improved the overall performance factor in compare to other classifier model. These steps help us in finding the correlation between variables and enable us to predict the classified output variable based on the predictor variables.