<![CDATA[This study aims to analyze and predict carbon dioxide (CO₂) concentration in parts per million (PPM) based on temperature and humidity using DHT22 and MQ135 sensors integrated with an ESP32 microcontroller as an Internet of Things (IoT) system. Data were collected in two measurement sessions. The first session was conducted from 16:00 to 20:00 with a one-second sampling interval, during which significant spikes occurred due to exposure to insecticide fumes. The second session was carried out from 00:00 to 08:00 with a two-second sampling interval, producing approximately 13,000 more stable data points suitable for machine learning analysis. The collected data were processed using several machine learning algorithms, including Linear Regression, Random Forest, CatBoost, LightGBM, XGBoost, and Neural Network (Keras). The results indicate that the CatBoost model achieved the best performance, with an accuracy of 98.9%, a Mean Absolute Error (MAE) of 2.75, and a Mean Relative Error (MRE) of 1.08%. This study demonstrates that an ESP32-based IoT implementation using integrated gas and temperature sensors can be utilized as a simple air quality prediction system and serve as a foundation for developing future intelligent environmental monitoring devices.]]>
Copyrights © 2026
