Detection of Arhythmia for Ballistocardiagram Based on CNN + RF

Abstract

In order to solve the problem that the total accuracy of ballistocardiogram (BCG) signal in arrhythmia detection is not high, the synchronous electrocardiogram(ECG) and BCG clinical signals of patients with premature ventricular contraction (PVC) and atrial fibrillation (AF) were collected Taking ECG signal as the standard, an automatic arrhythmia recognition model is constructed by using the hybrid algorithm of convolutional neural network (CNN) and random forest (RF). The experimental results show that the overall accuracy of the algorithm can reach 97.2%, which provides a new idea for the early detection and timely screening of cardiovascular diseases under the condition of natural sleep.