Phuntsho ChodenThara SeesaardTanthip Eamsa-ArdChutintorn SriphrapradangTeerakiat KerdcharoenMahidol University2018-12-212019-03-142018-12-212019-03-142017-03-232017 9th International Conference on Knowledge and Smart Technology: Crunching Information of Everything, KST 2017. (2017), 178-1812-s2.0-85017502548https://repository.li.mahidol.ac.th/handle/20.500.14594/42350© 2017 IEEE. In this work, we fabricated six chemiresistive sensors, employed in a portable e-nose and performed tests with urine samples from two groups of population, namely type II diabetes and healthy subjects. To identify sensitivity and selectivity of chemiresistive gas sensors, the first test was performed towards five volatile organic compounds (VOCs) which are particularly found in human urine profiles and the second test with real urine samples from the volunteers. Principal component analysis (PCA) and cluster analysis (CA) applied to validate the obtained sensing response successfully spilt urinary volatile odors into two separate groups of diabetes and healthy status. A hypothesis testing (p-value approach) demonstrated that S3 and S4 (p0.05) responded specifically to the urine odors from diabetic patients and healthy subjects. Our findings suggest the possibility of using chemiresistive gas sensors in e-nose as an alternative diagnostic tool for diabetes detection through analysis of volatile urine odors.Mahidol UniversityComputer ScienceConference PaperSCOPUS10.1109/KST.2017.7886086