Scholarlink

With the acceleration of global industrialization and urbanization, air pollution problems have become increasingly serious, posing a huge threat to human health and the ecological environment. The rapid development of machine learning and deep learning provides powerful tools for dealing with complex prediction problems and has achieved remarkable results. This paper aims to predict air pollutant concentrations, especially time series data of nitrogen dioxide (NO2) and ozone (O3), using a gradient boosted decision tree (GBDT) regression model. We extract moving average and Fourier transform features from historical pollution data, combined with the GBDT model, to improve prediction accuracy. Research results show that the proposed method can effectively capture the temporal dynamic change trend of pollutant concentrations and provide more accurate short-term predictions. This method provides new technical support for environmental pollution monitoring and control, and also provides data support and scientific basis for formulating relevant policies.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Predicting Environmental Pollution with Gradient Boosting: Application of GBDT Regression Models in Time Series Models

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Yijie Li,
  • Wangdong Jiang,
  • Guang Sun

摘要

With the acceleration of global industrialization and urbanization, air pollution problems have become increasingly serious, posing a huge threat to human health and the ecological environment. The rapid development of machine learning and deep learning provides powerful tools for dealing with complex prediction problems and has achieved remarkable results. This paper aims to predict air pollutant concentrations, especially time series data of nitrogen dioxide (NO2) and ozone (O3), using a gradient boosted decision tree (GBDT) regression model. We extract moving average and Fourier transform features from historical pollution data, combined with the GBDT model, to improve prediction accuracy. Research results show that the proposed method can effectively capture the temporal dynamic change trend of pollutant concentrations and provide more accurate short-term predictions. This method provides new technical support for environmental pollution monitoring and control, and also provides data support and scientific basis for formulating relevant policies.