Machine learning-based urban noise appropriateness evaluation method and driving factor analysis

• Published in: PloS one 2024-12, Vol.19 (12), p.e0311571
• Main Authors: Teng, Jinlin, Zhang, Cheng, Gong, Huimin, Liu, Chunqing
• Format: Article
• Language: English
• Subjects: Acoustics; Algorithms; Analysis; Biology and Life Sciences; China; Cities; City noise; Classification; Computer and Information Sciences; Data mining; Decision making; Decision trees; Earth Sciences; Ecology and Environmental Sciences; Engineering and Technology; Environmental management; Environmental Monitoring - methods; Environmental protection; Environmental quality; Evaluation; Factor analysis; Factor Analysis, Statistical; Geospatial data; Humans; Land use; Learning algorithms; Machine Learning; Methods; Noise; Noise control; Noise levels; Noise measurement; Noise pollution; Physical Sciences; Research and Analysis Methods; Roads; Roads & highways; Social Sciences; Sound; Spatial data; Spatial distribution; Sustainable development; Traffic; Trends; Urban areas; Urban development; Urban environments; Urban noise; Vehicles
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0311571

The evaluation of urban noise suitability is crucial for urban environmental management. Efficient and cost-effective methods for obtaining noise distribution data are of great interest. This study introduces various machine learning methods and applies the Random Forest algorithm, which performed best, to investigate noise suitability in the central urban area of Nanchang City. The findings are as follows: 1.Machine learning algorithms can be effectively used for urban noise evaluation. The optimized model accurately reflects the noise suitability levels in Nanchang City. 2.The feature importance ranking reveals that population spatial distribution has the most significant impact on urban noise, followed by distance to water bodies and road network density. These three features significantly influence the assessment of urban noise suitability and should be prioritized in noise control measures. 3.The weakly suitable noise areas in Nanchang's central urban region are primarily concentrated on the east bank of the Ganjiang River, making this a key area for noise management. Overall, the Unsuitable, Slightly suitable, Moderately suitable, Relatively suitable, and Height suitable areas constitute 9.38%, 16.03%, 28.02%, 33.31%, and 13.25% of the central urban area, respectively. 4.The SHAP model identifies the top three features in terms of importance, showing that different values of feature variables have varying impacts on noise suitability. This study employs data mining concepts and machine learning techniques to provide an accurate and objective assessment of urban noise levels. The results offer scientific decision-making support for urban spatial planning and noise mitigation measures, ensuring the healthy and sustainable development of the urban environment.