Paper Title
Cardiorespiratory Disorders Detection Using Photoplethysmography

Abstract
Based on the extraction of time domain features of Photoplethysmography signals, this study proposed an application of feature selection based classifier on Photoplethysmography technique. Data for 144 participants included healthy and cardiovascular disease patients was collected with five types of cardiovascular diseases. A Two stages classification methodology was used to identify cardiovascular disease type. At the first stage, subjects were classified as being normal and abnormal. Subjects identified as abnormal were then fed into the second stage classifier to identify the specific type cardiovascular disease. Based on the subset of features selected by the classifier, 10-fold cross validation accuracy of 88.88% and 85.43% for the two stages were obtained by Naïve Bayes classifier, respectively. This study proves the usefulness of Photoplethysmography signal in detecting cardiovascular disease occurrence, with employing the minimum number of Photoplethysmography time domain features, along with demographic variables and some health status information. Index Terms - Cardiovascular Diseases, Data Mining, Feature Selection, Photoplethysmography.