Irritable Bowel Syndrome Detection Based on Pulse Signal and Support Vector Machine

Abstract

In this developed society, most diseases appear due to stress people encounter day to day. Irritable Bowel Syndrome (IBS) is investigated in this study. 29 normal and 9 IBS subjects are involved in this study. Novel Traditional Chinese medicine’s pulse signal measurement system and pulse features are used. Pulse signal harmonicas and regression decay parameters are developed. Further multi-features classification by support vector machine showed that the left hand 9th harmonics and right hand alpha decay parameter are two candidates to differentiate IBS subject. SVM results showed classification performance of 85.7% accuracy, 100 % specificity and 44.4% sensitivity in this system.