Study on the Classification of Abnormal pulse signals Based on Deep Neural Network

The pulse of human body is affected by heart and blood, and carries important information reflecting the state of human body. Deep learning has made a breakthrough in features extracting for complex signals. The purpose of this paper is to establish a deep neural network to classify and identify the collected pulse data before and after long-time work. Firstly, the abnormal value of the original pulse data is replaced with the value within the normal range, and then the pulse length sequence 30 is divided into a short sequence of 75 consecutive sampling points. Finally, a deep neural network model is established to input short pulse sequence and output the corresponding physiological state. After training, the final classification accuracy of advanced neural network is 0.79 on 3200 training sets and 0.78 on 800 test sets. The pulse data before and after long-time work were effectively classified.