Research on Urban Air Quality Prediction System Based on Improved Random Forest Modelling

• Published in: Ecological chemistry and engineering. S Vol. 32; no. 2; pp. 213 - 235
• Main Authors: Su, Xin, Xin, Yifang, Yu, Yuekang, Zhao, Yue
• Format: Journal Article
• Language: English
• Published: Opole Sciendo 01.06.2025; De Gruyter Poland
• Subjects: Air; Air pollution; Air quality; air quality prediction; Algorithms; Bayesian analysis; bayesian optimisation; Bayesian theory; Cities; Construction; Data collection; Decision making; Decision trees; Economic development; Environmental monitoring; Feature extraction; Meteorological data; Optimization; Outdoor air quality; Predictions; Probability theory; random forest algorithm; Real time; Smart cities; smart city; Sustainable development; Urban planning
• ISSN: 2084-4549; 1898-6196; 2084-4549
• DOI: 10.2478/eces-2025-0011

With the rapid development of the digital economy, China’s smart city construction is facing great opportunities, especially in the field of environmental monitoring, which is very important for the development of smart cities. In this study, an advanced urban air quality prediction system is proposed to improve the monitoring ability and support data-driven urban planning decision-making. The system integrates low-cost distributed sensors and communication modules for real-time data collection and transmission, and realises intelligent feature extraction of atmospheric pollutant concentration data and meteorological data. In this system, Bayesian optimised random forest algorithm is used for hyperparameter optimisation and model prediction, and the prediction of air quality index (AQI) has high accuracy and reliability. The experimental results show that compared with the traditional random forest method, the Bayesian optimisation random forest algorithm can be applied to practice more accurately. Through feature extraction, hyperparameter optimisation and AQI evaluation, the system has the ability to automatically find the best “input feature + hyperparameter + model evaluation” for urban air quality. This research will be helpful to develop effective environmental monitoring tools for smart cities, and provide beneficial help for the construction and sustainable development of smart cities.