Breath Analysis\nUsing Quartz Tuning Forks for Predicting\nBlood Glucose Levels Using Artificial Neural Networks

Diabetes Mellitus (DM), a widespread metabolic disorder,\nposes\nlifelong health implications, demanding timely diagnosis and cautious\nmonitoring for effective disease management. Traditional blood glucose\ntests are invasive and require medical expertise for intermittent\nchecking, motivating the investigation of alternative, noninvasive\nmethods. This study introduces an approach employing breath analysis\nthrough a set of 12 quartz tuning fork-based sensors enhanced using\nnanomaterials and dedicated artificial neural network (ANN) algorithms\nfor data interpretation. The breath analysis methodology involves\ncapturing unique breath signatures using the frequency-based sensor\narray. The accompanying neural network classification algorithm, customized\nfor the sensor data, enables precise classification of data from 245\nindividuals as diabetic, prediabetic, or healthy. A neural network\nregression algorithm predicted blood glucose values and was compared\nwith the actual values obtained from medical blood glucose measurement.\nThe clinical relevance of the predicted blood glucose has been examined\nusing error grids. The sensor array coupled with the ANN algorithm\ncan identify diabetic, prediabetic, and control samples with 97% test\naccuracy. Blood glucose was predicted using neural network regression\nwith a correlation coefficient of 0.89 and a mean square error of\n0.13.