Fuzzy Classifier for Mental Stress Estimation using ECG Statistical Parameters

Abstract

Mental Stress estimation is an important feature to be derived in health related diagnostic activity. It has been observed that the stress has a major effect on heart functioning. And therefore, ecg should be the major source of stress variation and can be analyzed in various ways in order to extract the effect of mental stress. In the presented work, the ecg is analyzed using the statistical parameters set (energy, entropy, power, standard deviation and covariance). The parameters are not directly computed form the ecg itself. The ecg is first decomposed to level-2 using BIOR-3.9 wavelet transform to reduce the dimensionality of the ecg sample size. The level-1 and level-2 parameters are used to derive the mental stress levels as normal (N), hyper-1 (H-1), hyper-2 (H-2), depression-1 (D-1) and depression-2 (D-2). On parameter analysis, it has been observed that the energy and entropy are the two parameters that show an effective variation in values when normal to depression or normal to hyper case is observed. Therefore, the energy and entropy values are used for rule making and learning of the system in order to derive the mental stress levels

Improvement of human identification accuracy by wavelet of peak-aligned ECG

“…Analysis of heart rate or ECG morphology is useful to detect heart disease. Other information related to the person's health can also be estimated from ECG, such as respiration [4], stress level [5], energy expenditure [6], activity intensity [7], and activity classification [8]. Thus, daily monitoring of ECG at home is important to assess the person's health.…”