Efficient Data-Mining Classification Approach For Ecg Data In Health Care Application

Abstract

Data mining enables classification of Electrocardiographic (ECG) signals of the heart for diagnosing many cardiac diseases. ECG signals often consist of unwanted noises, speckles and redundant features. An unwanted noise and redundant features always degrade the quality of ECG signal and may lead to loss of accuracy in classification technique. To overcome these challenges, we introduced Optimize Discrete Kernel Vector (ODKV) classifier with an impressive pre-processing in this paper. In order to remove the noises, image processing filter namely the Adaptive Notch Filters (ANF) are initially used to remove Power Line Interference from ECG Signals. Moreover, reducing the redundant features from the ECG signal plays a vital role in diagnosing the cardiac disease. So, Optimize Discrete Kernel Vector (ODKV) classifier is used to reduce the redundant features and also to enhance the classification accuracy of the input ECG signal. Thus, Optimize Discrete Kernel Vector (ODKV) classifier identifies the Q wave, R wave and S wave in the input ECG signal. Finally, performance metrics Sensitivity, Specificity, Accuracy and Mean Square Error (MSE) are calculated and compared with the existing method such as SVM-kNN, ANN-kNN, GB-SVNN, and CNN to prove the enhancement of the classification technique.