Fatigue and emotion analysis method based on ECG signal

The invention discloses a fatigue and emotion analysis method based on ECG signal. The method is based on PCA, K-means clustering algorithm and a realization of DS evidence theory. The method specifically comprises the following steps: collecting the ECG signal data, performing noise detection, weak signal detection and filtering processing on the ECG signal data, extracting R-R interval sequencesof the ECG signal data; analyzing a time-frequency domain of the R-R interval sequence to obtain a time-frequency domain index for data training, and carrying out interval division on the acquired time-frequency domain index through using the DS evidence theory; carrying out the PCA on the time-frequency domain indexes of the interval division through the DS evidence theory to acquire a low-dimensional matrix and a corresponding transformational matrix; carrying out clustering processing on the acquired low-dimensional matrix through the K-means clustering algorithm, dividing the categories corresponding to different