Design and testing of electronic nose for determining the pattern of bad breath classification in patients with diabetes mellitus and pulmonary tuberculosis (TBC)

In this study, the electronic nose (E-Nose) is based on array gas sensor. This tool consists of 10 MQ gas sensors related to Volatile Organic Compound (VOC) gas components. The working principle of the tool is to imitate the biological system of the nose which is useful to smell both bad and pleasant smell. E-Nose will detect samples of breath or bad breath on Tuberculosis (TBC) and Diabetes Mellitus patients and healthy people. Then, the data responses will be evaluated by using 2 patterns recognition methods called Principle Component Analysis (PCA) and Cluster Analysis(CA). The classification results can be explained by the value of the first 3 PCs from the score plot on the PCA of the data. PC1 accounts for 52.6% of the variance, while PC2 accounts for 17.8% of the variance, and PC3 accounts for 9.5% of the variance. The cumulative value of the first 3 PCs is 79.9%. The score plot graph shows a perfect classification of 3 data groups between healthy people, tuberculosis and diabetes mellitus patients. The results of this study indicate that the electronic nose (E-Nose). The instrument can identify the pattern of bad breath of healthy people and TBC and diabetes mellitus patients