The emotional state classification using physiological signal interpretation framework

This paper proposes and evaluates an emotional state classification using a physiological signal interpretation framework. The proposed Emo-CSI framework consists of three components which are the following: 1) physiological signal sensing, 2) data pre-processing, and 3) emotional state classification. The Emo-CSI framework applies physiological signals (i.e., heart rate, breathing pattern, skin temperature, skin humidity, and skin conductivity) to classify the emotional state. The emotional state classification results in an emotional state (i.e., displeasure, neutral, pleasure, calm, medium, and excited). This research also investigates the accuracy of three classification techniques which are the following: 1) support vector machine (SVM), 2) artificial neural network (ANN), and 3) decision tree (DT). The evaluation results show that the physiological signals are related to emotional state. Using SVM as a classification in Emo-CSI outperforms the other classification techniques.