K-means Clustering Algorithm for Myocardial Infarction Classification

Authors: ElAmin A, Alotaibi YM, Abaida EA, Malaekah E, Ismail Saied HF* and Mukhanov VV

DOI: 10.23880/aabsc-16000113

Volume 2, Issue 1

Abstract

Heart attack is one of the main causes of death around the world. The electrocardiogram (ECG) is considered as an effective method to diagnose heart diseases. In this study, the classification approach will be applied to distinguish between myocardial infarctions (MI) subtypes. The abnormalities approval depends on morphological analysis used to detect the appearance or absence of some specific features in ECG graph. The classification approach consists of many basic steps such as data segmentation and various kinds of noise removal. Stemming filter will be used for detection of some features through a large number of data. Features selection and normalization approaches will be used for data validation and significant clustering. Finally, the K-means clustering will be used to differentiate between MI subtypes. The statistical analyses such as Precision, Recall, and F-Score were used to evaluate the performance of the K-mean algorithm. The F-score achievement is indicated at 0.98% in significant clustering and with 87.61% in right classification of MI subtypes.

Keywords: Myocardial Infarction; K-means Clustering; Stemming Algorithm; Features Extraction; Morphological Root

View PDF