Negative Emotion Detection using ECG and HRV Features

Emotions cause different physical, behavioural and cognitive changes in the human body. Emotions can be positive and negative. Negative emotion is the experience of negative feelings such as anger, frustration, panic, stress and fear. These negative emotions can cause severe health problems. So there is a need for detection of negative emotions . It will help in improving the health of the human body. As these emotions result in a change of various physiological parameters like heart rate, skin temperature, blood pressure, skin conductance, etc., these signals can be used to detect the emotions of a person. These signals are generated by the body during the functioning of various physiological systems, so they cannot be regulated artificially. Due to this reason, it is a reliable source for the detection of such information. So physiological signal is one of the most important factor in the field of emotion detection. The change in signals represents certain characteristics which are used to estimate the emotions. This work mainly focuses to build a better model of negative emotion detection for Typically Developed group using Machine learning approach with the help of Electrocardiogram (ECG) signal. This study was conducted on DECAF database for typically developed group. The study focused to extract the relevant features from both ECG and HRV signals. Then to identify which is more contributing towards negative emotion detection. A machine learning model was developed for typically developed group db4 as mother wavelets for feature extraction. The significant features of ECG and HRV were then classified separately using the logistic regression, ensemble and support vector machine. Logistic regression classifier achieved maximum accuracy using HRV data for typically developed (TD) group