Breath Acetone Based Noninvasive Blood Glucose Monitoring System for Diabetes Using SVM

A novel method for monitoring blood glucose level based on the respective change in the acetone concentration of the diabetic patient is designed and tested with laboratory prepared acetone gas samples. The results discussed here are impressive and promise a possibility of non-invasive breath acetone based blood Glucose monitoring system which reduces the number of blood samples to be analysed each day for precise monitoring of blood glucose level. Support Vector Machine a powerful regression tool is used for principal component analysis. A Volatile Organic Component sensor by FIGARO is used to monitor the acetone concentration in the simulated gas samples. The classification is done in to two groups and the results are visually interpreted from the graph generated by the computer that is integrated with the data acquisition unit using a serial communication port. Sensing element in is a thin strip of tin oxide.