Performance Assessment of Machine Learning Classifiers for Detecting Heart Rate Variations from PPG Signals

This paper is going to discuss about the detection of heart disorder due to variation of heart rate, and also analyze the different classifiers performance in the finding of the disorder. The five statistical features like mean, skewness, variance, kurtosis and the correlation of PPG signal is considered in this work of PO2 measurement. The PPG features will indirectly indicate the condition of heart rate .The classifiers such as SDC, Firefly, K-Nearest Neighbor (KNN), Kernel PCA, Gaussian Mixture Model (GMM), PCA Firefly are used in the classification of normal and abnormal heart rate disorder from PPG signal feature. The parameters accuracy, specificity, sensitivity, Error Rate, PI are used to analyze the performance of the classifiers. The classifier that obtained superior performance is the focus of the findings, as it indicates its effectiveness in comparison to other classifiers.