Scholarlink

<p>For environmental monitoring and making decisions that are good for the environment, it is important to be able to accurately anticipate CO<InlineEquation ID="IEq3"><EquationSource Format="TEX">\(_2\)</EquationSource></InlineEquation> emissions. This paper presents an interpretable deep learning framework utilizing a BiLSTM network combined with SHAP for CO<InlineEquation ID="IEq4"><EquationSource Format="TEX">\(_2\)</EquationSource></InlineEquation> prediction. Using a sliding window technique, CO<InlineEquation ID="IEq5"><EquationSource Format="TEX">\(_2\)</EquationSource></InlineEquation> time-series data are reformed to show how things change over time. To stop data leaking, chronological data separation and min-max normalization are used. The proposed BiLSTM model achieved the best result. It has an RMSE of <InlineEquation ID="IEq6"><EquationSource Format="TEX">\(8.58 \times 10^{5}\)</EquationSource></InlineEquation>, an MAE of <InlineEquation ID="IEq7"><EquationSource Format="TEX">\(1.19 \times 10^{5}\)</EquationSource></InlineEquation>, a <InlineEquation ID="IEq8"><EquationSource Format="TEX">\(R^2\)</EquationSource></InlineEquation> score of 0.99, and a MAPE of 0.0073% on an unseen dataset. Global and local explanations based on SHAP give complementing feature-level insights using a distinct explainability model trained on engineered country, sector and temporal attributes.</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

A scalable explainable deep learning framework for predictive analysis and interpretability of CO\(_2\) emissions patterns

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月31日

查看全文
  • Debyanshu Tiwari,
  • Sandeep Saharan,
  • Deepanshu Kaushik,
  • Muzafar Ahmad Wani,
  • Rajesh Kumar Chaudhary,
  • Jatin Bedi,
  • Niyaz Ahmad Wani,
  • Mudasir Mohd,
  • Pinky Yadav

摘要

For environmental monitoring and making decisions that are good for the environment, it is important to be able to accurately anticipate CO\(_2\) emissions. This paper presents an interpretable deep learning framework utilizing a BiLSTM network combined with SHAP for CO\(_2\) prediction. Using a sliding window technique, CO\(_2\) time-series data are reformed to show how things change over time. To stop data leaking, chronological data separation and min-max normalization are used. The proposed BiLSTM model achieved the best result. It has an RMSE of \(8.58 \times 10^{5}\), an MAE of \(1.19 \times 10^{5}\), a \(R^2\) score of 0.99, and a MAPE of 0.0073% on an unseen dataset. Global and local explanations based on SHAP give complementing feature-level insights using a distinct explainability model trained on engineered country, sector and temporal attributes.