首页|贝叶斯优化模糊聚类地级行政区声环境

贝叶斯优化模糊聚类地级行政区声环境

扫码查看
声环境功能区划是噪声污染管理的重要手段.当前声环境功能区划研究大多是基于某个特定的地级行政区来进行的,难以反映各地级行政区声环境的异同.该文基于134个地级行政区的人口、面积、各声环境功能区面积和及面积占比,进行地级行政区声环境表征和归一化处理.以轮廓系数作为聚类有效性评价指标,基于贝叶斯优化模糊聚类方法对地级行政区声环境表征进行聚类分析.通过与谱聚类、K-medoids聚类、高斯混合模型聚类的聚类性能对比,验证了该方法的有效性.结果表明,我国地级行政区声环境分为9类,城市规模和用地情况发生显著变化导致其声环境表征和归类发生变化后应重新评估当前噪声污染管理政策,并借鉴同类的地级行政区的噪声污染管理政策做出必要的调整.
Bayesian optimal fuzzy clustering of acoustic environment in prefecture-level administrative regions
Environmental noise function zoning plays a crucial role in noise pollution management.Recent re-searches on environmental noise function zoning were mainly based on a specific prefecture-level administrative region and did not reflect the similarities and differences between various prefecture-level administrative regions.In the paper,the acoustic environment in 134 prefecture-level administrative regions was characterized and normalized based on the administrative region's population,area,and environmental noise functional zones.Cluster analysis of the acoustic environment of prefecture-level administrative regions was conducted based on Silhouette criterion clustering evaluation and Bayesian optimization fuzzy clustering.This method is verified by comparing its clustering performance with spectral clustering,K-medoids clustering,and Gaussian mixture model clustering.Results indicate that the acoustic environment of prefecture-level administrative regions are divided into nine categories.The current noise pollution management policies should be re-evaluated and ad-justed as necessary after significant changes in urban size and land use lead to changes in the characterization and classification of acoustic environment in a prefecture-level administrative region.

Environmental noise function zonePrefecture-level administrative regionFuzzy clusteringBayesian optimizationSilhouette coefficient

曾宇、姚琨、任爽、户文成

展开 >

北京市科学技术研究院城市安全与环境科学研究所 北京 100054

声环境功能区 地级行政区 模糊聚类 贝叶斯优化 轮廓系数

2024

应用声学
中国科学院声学研究所

应用声学

CSTPCD北大核心
影响因子:1.128
ISSN:1000-310X
年,卷(期):2024.43(2)
  • 30