AI‐Enabled Hematene‐Based Nitric Oxide Gas Sensor Suitable for Respiratory Disease Monitoring

Abstract This study presents the development of a hematene‐based gas sensor integrated with machine learning techniques for accurate detection of exhaled nitric oxide (NO), a key biomarker for respiratory diseases such as asthma and chronic obstructive pulmonary disease (COPD). The sensor exhibits a response of 89.9% to NO at 150 °C with a sensitivity of 38.6 kΩ/ppb in the 20–400 ppb range and a limit of detection of ≈401 ppt. The response and recovery times are 23 and 118 s, respectively, with high repeatability. Although the sensor exhibits good selectivity at low concentrations, accurately identifying NO gas in the presence of interfering gases such as ammonia, nitrogen dioxide, and acetone in the mixed environments remained challenging. To address this, various machine learning models like Random Forest, Support Vector Machine, CatBoost, Neural Networks, etc., are explored. The RF model achieved the highest classification accuracy, 98.6% for individual gases, 97.5% for binary mixtures, and 89.7% for the ternary mixture, along with a 23.91% improvement in NO gas selectivity via regression analysis. Principal Component Analysis is also used for visualization to classify gas types distinctly. This AI‐assisted sensing framework demonstrates strong potential for non‐invasive, real‐time, and cost‐effective in healthcare monitoring applications.