DESIGN AND PERFORMANCE ANALYSIS OF MICRO-CANTILEVER BASED AMMONIA GAS SENSOR USING MACHINE LEARNING

Ammonia gas is a biomarker for different clinical conditions, when it exceeds the normal concentration in the exhaled breath. The detection of ammonia in humans helps to
investigate various practicalities about kidneys, liver and bacterial infection of either the stomach or mouth. The Micro-cantilever-based ammonia gas sensor is designed in COMSOL Multiphysics 5.0V. This data is procured from the simulation result and predictive machine learning models
have been used to assess the accuracy of the data being sensed. In this project, we demonstrate the best machine learning algorithms like SVR (Support Vector Regression), Linear Regression, Random Forest etc. for our performance analysis and each algorithm is compared with other algorithms. This study illustrates that machine learning has respectable potential for the prediction of ammonia levels which can be further used to predict the various clinical
conditions. Future research should consider boosting, or using nature inspired to develop a gastric cancer prediction model.
Keywords: Biosensor, Micro-cantilever, COMSOL Multiphysics 5.0V, Support Vector Machine, Linear Regression