Abstract:In order to improve the recognition rate of arrhythmia automatic diagnosis, we proposed a classification method of
arrhythmia based on XGBoost model. The feature vector of MLII lead signal was composed by signal projection feature and
RR interval feature. After standardized processing of feature vector, it was inputted to the XGBoost classifier. According to the
AAMI standard, the XGBoost-based arrhythmia classification model can classify the ECG signals into four categories: normal
beat (N), supraventricular ectopic beat (S), ventricular ectopic beat (V) and fusion beat (F), respectively. The MIT-BIH
arrhythmia database was employed in model test. Comparison with the existing arrhythmia detection method using the case of
the same data set, the average recognition accuracy of XGBoost model was further improved to 94.1%.
李云,吴水才,袁丽. 基于XGBoost模型的心律失常分类算法研究[J]. 中国医疗设备, 2019, 34(7): 24-28.
LI Yun, WU Shuicai, YUAN Li. Research on Arrhythmia Classification Algorithm Based on XGBoost Model. China Medical Devices, 2019, 34(7): 24-28.