ECG signal diagnosis using Discrete Wavelet Transform and K-Nearest Neighbor classifier.

Early detection of heart problems can save the lives of many people around the world; the electrical activity of the heart represented by electrocardiograms (ECGs) gives us important information about the health of the heart and it useful tool to detect certain cardiovascular diseases. Efficient analysis of ECG signals to extract features can be a good tool to diagnosis of these diseases; this features help us to identify the ECG signal and make the good decision in the diagnosis. The purpose of this work is to process the ECG signals using the wavelet analysis; this tool allows in the same time to denoising the signal and better locate any abnormalities that this signal presents. The features used for the characterization of the signal are obtained by extracting the statistical features from wavelet coefficients. The distinction between the signals is made by a classification of the signals, among the different existing classification methods we have adopted in this study the K Nearest Neighbors KNN method .The use of the discrete wavelet transform DWT with the Symlet 8 as a mother wavelet and the K Nearest Neighbors classifier allowed us to establish a model which is used to identify these ECG signals with an accuracy which reaches 91.60%.