According to the diversity, complexity and how to identify nonlinear mapping of the QRS waveform, this paper proposes an improved BP(Back Propagation) algorithm. In the detection and recognition of QRS, after completed extraction for electrocardiogram signal feature point, this algorithm is used for rapid training analysis. Several typical and abnormal QRS waves have been automatic recognized and classified successfully. Detailed simulation experiments show that after the algorithm applied to electrocardiogram diagnosis, it has a good classification effect and faster training speed.
展开▼