No abstract

No abstract

No abstract

This study describe a feature extraction method based on linear prediction for classification of QRS in an associative memory model.The feature extraction will convert each QRS pattern to a pulse-code-train which describes only -1 , 0 , and + 1 three states. In order to recognize the feature for QRS pattern, we provides a two-layer forward connecting neural nets model in this study. The model operates each input node as well as a real neuron's three typical states; resting state [ O ] , excitatory state [ + I I , and inhibitary state [-1].We present this model's manner is all the same as others and the classification rate performs high efficiency for arrhythmia detection.