Simplified support vector machine method for QRS wave detection

Recently, support vector machine (SVM) has been applied to ECG wave detection fields successfully, based on its salient properties that other learning method do not provide. One drawback for SVM is that the time taken for classifying a new pattern is proportional to the number of support vectors, so if that number is large, classification speed is slow, which will impair the detection speed of QRS wave. To overcome the problem, a novel QRS wave detection model relying on simplified SVM is presented in this paper, which can improve classification speed greatly by compressing the number of support vectors. As a result, the real-time response capability of QRS wave detection has been improved, significantly.