A Novel Artificial Intelligence-Enabled Method for Electronic Nose Design Based on Olfactometry Data

Electronic nose systems are advanced technological tools that enable the objective evaluation of odors through sensor arrays mimicking the human olfactory mechanism and sophisticated data processing algorithms. These systems facilitate rapid, reproducible, and standardized measurement of chemical components in applications such as food safety, environmental monitoring, medical diagnostics, and industrial quality control. In this study, measurements obtained from electronic nose sensors were compared with olfactometry panelist assessments using n-butanol as a reference substance in accordance with the TS EN 13725 standard. Furthermore, machine learning algorithms, including Partial Least Squares (PLS), Support Vector Regression (SVR), and Gaussian Process Regression (GPR), were applied to model the sensor data and evaluate their predictive accuracy. The results demonstrated the reliability and applicability of the electronic nose system, achieving training mean absolute percentage error (MAPE) values of 6.53% for PLS, 10.89% for SVR, and 0.15% for GPR. This study presents an innovative approach that systematically assesses the performance of electronic nose technology using a standardized reference odor and highlights the effectiveness of the modeling approach.