Scholarlink

<p>Air quality significantly impacts public health, industrial stability, and timely responses to environmental hazards, all of which are essential for sustainable development. Accurate forecasting of the Air Quality Index (AQI) is therefore crucial for effective environmental monitoring and management. In this study, we develop a hybrid deep learning model that integrates a Transformer encoder with a Bidirectional Long Short-Term Memory (BiLSTM) network. The model is trained and validated using daily air quality data collected from Shijiazhuang, Beijing and Tianjin, spanning November 2013 to February 2025. Experimental results demonstrate that the proposed Transformer-BiLSTM model delivers stable and reliable predictive performance, with root mean squared error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE) of 3.0012 ug/m<InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(^3\)</EquationSource> </InlineEquation>, 1.7928 ug/m<InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(^3\)</EquationSource> </InlineEquation>, and 3.3646%, respectively. Compared with conventional baseline models, the hybrid model improves accuracy and generalization capability. This approach offers a reliable and interpretable tool for AQI forecasting and provides quantitative support for data-driven air pollution control strategies.</p>

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

Intelligent prediction of air quality index based on the transformer-BiLSTM model

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年11月25日

查看全文
  • Xinni Liu,
  • Kai Su,
  • Shubin Wang,
  • Kamarul Hawari Ghazali

摘要

Air quality significantly impacts public health, industrial stability, and timely responses to environmental hazards, all of which are essential for sustainable development. Accurate forecasting of the Air Quality Index (AQI) is therefore crucial for effective environmental monitoring and management. In this study, we develop a hybrid deep learning model that integrates a Transformer encoder with a Bidirectional Long Short-Term Memory (BiLSTM) network. The model is trained and validated using daily air quality data collected from Shijiazhuang, Beijing and Tianjin, spanning November 2013 to February 2025. Experimental results demonstrate that the proposed Transformer-BiLSTM model delivers stable and reliable predictive performance, with root mean squared error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE) of 3.0012 ug/m \(^3\) , 1.7928 ug/m \(^3\) , and 3.3646%, respectively. Compared with conventional baseline models, the hybrid model improves accuracy and generalization capability. This approach offers a reliable and interpretable tool for AQI forecasting and provides quantitative support for data-driven air pollution control strategies.