부산대학교 도서관

Observing the condition of the cardiovascular system is a vital task in the medical sector. The electrocardiogram (ECG) is such a tool that can be used to detect cardiovascular abnormalities. The advanced techniques of Machine Learning can help us to detect such abnormalities with the help of computers. But to effectively train the machine, we need to extract meaningful features from the ECG signals instead of using the raw signal as input. In this study, a set of handcrafted features have been extracted after signal preprocessing and used to train a classifier properly. The aim of this paper is to propose an effective technique to classify 17 different classes of ECG signals based on an ensemble learning algorithm named Random Forest (RF) classifier. The method provides 88% classification accuracy.