Preliminary abnormal electrocardiogram segment screening method for Holter data based on long short-term memory networks
Chen Siying ‡, ‡, Liu Hongxing
       

Kernel density estimation and confidence interval determination: (a) histogram of the prediction error produced by the prediction model on the normal signal points, (b) kernel density estimation based probability density function curve of normal ECG signal points, (c) confidence threshold and determined corresponding left and right thresholds through integration.