Detection of abnormal heart conditions from the analysis of ECG signals

Actual classification of Electrocardiogram (ECG) signals is vital and necessary for detection of abnormal heart conditions from the analysis of ECG signals. In this paper, we have proposed a classifier that simulates the diagnosis of the cardiologist to classify ECG signals data into two classes: normal and abnormal classes from single lead ECG signals and better than other well-known classifiers. The proposed classifier solved most of well-known classifiers problems also, overcomes the misdiagnosis problems that face many cardiologists. The proposed algorithm is validated using 48 records from the MIT-BIH arrhythmia database, where 25 records for normal class and 23 records for abnormal class. Two Neural Network (NN) classifiers: Feed Forward Network (FFN) and Multi-layered Perceptron (MLP), four Support Vector Machine (SVM) classifiers: Linear-SVM, Gaussian Radial Base function (RBF), Polynomial-SVM and Quadratic-SVM and K-Nearest Neighbour (KNN) classifier are employed to classify the ECG signals and compared with the proposed classifier. The total 13 features extracted from each ECG signal used in the proposed algorithm and set as input to the other classifiers. Experimental results show that the proposed classifier demonstrates better performance than other classifiers in terms of accuracy and computing time.