ECG signal classification using Hjorth Descriptor

ECG signal occurs due to heart´s electrical activity and helps detect and record people´s heart health. Many methods have been developed to classify ECG signal automatically. In this research, Hjorth Descriptor is used as a method for feature extraction. K-Nearest Neighbor (KNN) and Multilayer Perceptron (MLP) are used as classifier in classification stage. Experiment result showed that both K-NN and MLP achieved accuracy up to 100% for 50% of test data. Results of 99.33% accuracy were obtained for 10-fold cross validation. Hence, Hjorth Descriptor generates a good feature related to ECG signal classification process.