COMPARATIVE ANALYSIS OF MACHINE LEARNING ALGORITHMS FOR CO CONCENTRATION MONITORING USING IOT SENSORS

Abstract

This paper presents a comparative analysis of machine learning (ML) algorithms for monitoring carbon monoxide (CO) concentration using IoT sensors (SENSOR - Mine 4GN) in confined industrial spaces. The study considers linear regression, ensemble models such as Random Forest and XGBoost, support vector machines (SVM) and the recurrent neural network LSTM. It is shown that LSTM provides the highest prediction accuracy and minimizes delays in detecting abnormal CO spikes. The proposed methodology combines initial noise filtering, alarm confirmation rules, and ML-based forecasting, which enhances the reliability of the monitoring system and reduces false alarms. The results confirm the effectiveness of integrating IoT sensors with intelligent data processing for safety management in mines and underground facilities.