Prediction of Acute Coronary Syndrome Using Pulse Plethysmograph

Acute Coronary Syndrome (ACS) is one the major reason of increasing mortality rate. Despite medical advances, there isn't any efficient method that can control this proliferating mortality rate. The major aim of this study is detection of ACS using Pulse Plethysmograph (PuPG) signal analysis. 348 samples of PuPG signal were acquired by fastening PuPG sensor to subject's index finger. Signal data is preprocessed through Empirical Mode Decomposition (EMD) to remove any possible noise and other artifacts. Extensive experimental analysis was performed to select features having maximum intraclass distance and discriminative power to classify ACS and Normal signals through Support Vector Machines (SVM) classifier. 5-fold cross validation was used to perform training and testing of proposed model using self-collected dataset of PuPG signals. Average accuracy of 99.42%, sensitivity of 99.43% and specificity of 99.41% is obtained through SVM with cubic kernel proving proposed model as a best possible methodology in-terms of cost and efficiency as compared to existing solutions.