Arrhythmia heart beats classification using mahalanobis Generalized Learning Vector Quantization (Mahalanobis GLVQ)

Automatic heart beats classification has attracted much interest for research recently and we are interested to determine the type of arrhythmia from electrocardiogram (ECG) signal automatically. In this study we modified supervised Generalized Learning Vector Quantization (GLVQ) A. Sato with inject the mahalanobis distance to GLVQ in order to develop a robust algorithm that we said Mahalanobis GLVQ. The overall classification system is comprised of three components including data pre-processing, feature extraction and classification. Data preprocessing related to how the initial data prepared, while for the feature extraction and selection, we using wavelet algorithm. The classification will be divided into two phases, at ones phase we will test the algorithm with clean data from outlier, and at second phase we use noisy data that contains outlier data. The ECG signals are obtained from MIT-BIH arrhythmia database. Accuracy of Mahalanobis GLVQ in our study is 92% for clean data test with 24 feature and 87% for GLVQ, its show that Mahalanobis GLVQ able to increasing accuracy of GLVQ. The result experiment of data test that contain outlier data, accuracy of Mahalanobis GLVQ in our study is 67% and accuracy or GLVQ is 65%.