A hybrid case-based reasoning approach for the electrocardiogram diagnosis

The death caused by heart diseases has become a serious problem, how to diagnose heart diseases efficiently plays a more important role recently. Because of the high development of computer technologies, using the Artificial Intelligence approaches to analyze and diagnose ECG is an important topic. This paper proposes an ECG automatic diagnosis system based on the Case-Based Reasoning (CBR). First, at the pre-processing stage, we use the bilateral filter to remove noise. At the feature extraction stage, the ECG feature points are detected by the moving average method and the differential equation approach. Finally, at the recognition stage, we propose a new hybrid CBR architecture, which combines the Artificial Neural Network (ANN) with the traditional CBR. We cluster and index the cases by the ANN, and then use the weighted Euclidean distance as the similarity measurement technique for case retrieving. We also provide an interactive user interface for allowing the experts to adapt the retrieved case. From our experiment, it shows that the proposed system is a good automatic diagnosis system of ECG which has the following advantages, high clustering performance and the quick case retrieving time.