Certain Examination on ECG Classification using Hybrid Machine Learning Models

A Non-stationary signal called an electrocardiogram (ECG) is utilized to evaluate the rate of cardiac beats. Doctors can identify Abstract - Heart Rate Variations (HRV) is one of the major issues focused in Medical field. The unexpected functionality of heart is measured from the Electrocardiogram (ECG) signals. The HRV is one of severe disease which can affect any people of any age and it cause to sudden death. This research article focusses on identifying the HRV using the various Machine Learning (ML) algorithms. This research work develops the hybrid ML (HyML) which has better performance in all the aspects. The HyML is constructed with Support Vector Machine (SVM), Robust principal Component Analysis (RPCA) and Hidden Markov Models (HMM). Also, to support the different data of ECG multi SVM (MSVM) and Single SVM (SSVM) models were developed. These models are trained with 5348 samples under 8 different categories to classify the HRV. Therefore we propose the model HyML, which improvise the effectiveness on classifying the HRV. Our proposed model achieves the accuracy of 94.S3% for MSVM and 93.S9% for SSVM. The overall the sensitivity is varying between 0. S2-0.S9 for SSVM and 0.S2-0.9l for MSVM. Also, the precision observed in the range of 0. S7-0.91 for SSVM and 0.91 - 0.94 for MSVM. Hence, the proposed model is highly recommended model for HRV classification with SVM of both models such as single and multi.