Smart sustainability: Environmental accounting strategy for modern corporations using machine learning

With increasing climate change urgency and regulatory pressures, corporations are expected to adopt transparent and efficient environmental accounting practices. In China, where industrialization often outpaces sustainability, there is a need for intelligent tools to align business performance with environmental responsibility. Despite the existence of environmental key performance indicators (KPIs), many organizations struggle with adopting suitable environmental accounting strategies due to a lack of data-driven frameworks. Traditional approaches often overlook the complexity and industry-specific nature of sustainability data, leading to poor decision-making. This research proposes EcoStratClass, a machine learning-based framework for classifying Chinese corporations into appropriate Environmental Accounting Strategy types based on sustainability performance. Using the Smart Sustainability & Environmental Accounting in Chinese Corporations (SSEC-ChiCorp) dataset, the methodology involves advanced data preprocessing, including attribute removal, target mean encoding, PCA for categorical reduction, ordinal mapping for audit frequency, and hybrid normalization techniques (Z-score, Min-Max, and Box-Cox). The Environmental Signature Score (ESS) quantifies overall sustainability impact, considering CO 2 emissions, energy usage, recycling, and renewable resource consumption. Feature selection employs the Environmental Signature Learning (ESL) method, combining Mutual Information, Chi-Square, ANOVA F-test, Recursive Feature Elimination (RFE), and models like LightGBM, Random Forest, and SVM. The classification model uses a stacked ensemble of LightGBM, Random Forest, and SVM, with Logistic Regression as the meta-learner. SHAP values improve model explainability by highlighting influential attributes. With an accuracy of 92.87%, the model achieves strong performance in various metrics (macro precision, recall, F1 score). EcoStratClass offers a reliable, interpretable decision-support system to guide corporations in selecting effective environmental accounting strategies, promoting ESG-aligned sustainable development.

Authors

Institutions

Publication Details

Journal
Intelligent Decision Technologies
Published
2025-07-14
DOI
https://doi.org/10.1177/18724981251357820
Citations
2
Primary Topic
Environmental Sustainability in Business
Type
article
Field-Weighted Citation Impact
4.15
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Smart sustainability: Environmental accounting strategy for modern corporations using machine learning

Junping Ren, H. J. Yang, Zhao Wen-hua
2 citations
Intelligent Decision Technologies
Environmental Sustainability in Business
4.15
article

Smart sustainability: Environmental accounting strategy for modern corporations using machine learning

Junping Ren, H. J. Yang, Zhao Wen-hua
article en
2 citations

Abstract

With increasing climate change urgency and regulatory pressures, corporations are expected to adopt transparent and efficient environmental accounting practices. In China, where industrialization often outpaces sustainability, there is a need for intelligent tools to align business performance with environmental responsibility. Despite the existence of environmental key performance indicators (KPIs), many organizations struggle with adopting suitable environmental accounting strategies due to a lack of data-driven frameworks. Traditional approaches often overlook the complexity and industry-specific nature of sustainability data, leading to poor decision-making. This research proposes EcoStratClass, a machine learning-based framework for classifying Chinese corporations into appropriate Environmental Accounting Strategy types based on sustainability performance. Using the Smart Sustainability & Environmental Accounting in Chinese Corporations (SSEC-ChiCorp) dataset, the methodology involves advanced data preprocessing, including attribute removal, target mean encoding, PCA for categorical reduction, ordinal mapping for audit frequency, and hybrid normalization techniques (Z-score, Min-Max, and Box-Cox). The Environmental Signature Score (ESS) quantifies overall sustainability impact, considering CO 2 emissions, energy usage, recycling, and renewable resource consumption. Feature selection employs the Environmental Signature Learning (ESL) method, combining Mutual Information, Chi-Square, ANOVA F-test, Recursive Feature Elimination (RFE), and models like LightGBM, Random Forest, and SVM. The classification model uses a stacked ensemble of LightGBM, Random Forest, and SVM, with Logistic Regression as the meta-learner. SHAP values improve model explainability by highlighting influential attributes. With an accuracy of 92.87%, the model achieves strong performance in various metrics (macro precision, recall, F1 score). EcoStratClass offers a reliable, interpretable decision-support system to guide corporations in selecting effective environmental accounting strategies, promoting ESG-aligned sustainable development.

Intelligent Decision TechnologiesVol. 19(5)
Anhui Agricultural University (CN), Anhui University (CN), Shaanxi University of Science and Technology (CN)
Openalex Percentile: Top 6%
Environmental Sustainability in Business
4.15
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.