Optimization of Gas Sensor Array in Electronic Nose Using Whale Optimization Algorithm for Tuberculosis Detection

A lung infected with tuberculosis contains volatile organic compounds (VOCs) caused by Mycobacterium tuberculosis. An exhaled breath analysis approach can detect biomarkers of tuberculosis. An electronic nose can be proposed as a device for the early diagnosis of tuberculosis. The primary problem with this device is to specify the optimal number of gas sensors while maintaining high performance. The whale optimization algorithm (WOA) is an optimization method that uses the hunting mechanism of humpback whales, including bubble-net feeding and spiral movement. This algorithm is straightforward to implement, uncomplicated, and converges quickly. This study aimed to determine the optimal number of gas sensors in an electronic nose for differentiating between patients with tuberculosis and healthy controls through exhaled breath using WOA. The implementation was performed using Python programming and Jupyter Notebook for data processing, optimization, and classification. The experimental results demonstrate that the WOA efficiently selected the optimal gas sensor combinations in a short optimization time. Furthermore, two gas sensors were selected, and the support vector machine (SVM) successfully distinguished between tuberculosis and healthy controls with accuracy, precision, recall, and f1-score above 98%.