Scholarlink

<p>Environmental, Social, and Governance (ESG) practices are increasingly used to quantify sustainable corporate behavior and are often linked to long-term financial performance and resilience in public equity markets. This study proposes an AI-based predictive framework to analyze the relationship between ESG-related indicators and firm performance using the Kaggle 400k NYSE Random Investments + Financial Ratios dataset, enriched with additional financial and macroeconomic variables. The data are standardized via Z-score normalization, compressed using Principal Component Analysis (PCA), and refined through Recursive Feature Elimination (RFE) to retain the most informative predictors. For prediction, we develop an Optimized Attentive Ant Colony Convolve Neuro Recurrent Net model (ACO-CNRC) that integrates Ant Colony Optimization for configuration search with CNN/RNN sequential modeling and an attention mechanism for improved representation learning. Experiments on the NYSE-based dataset achieve high predictive performance, reaching 99.74% accuracy on the classification task and, <InlineEquation ID="IEq1"><EquationSource Format="TEX">\({R}^{2}\)</EquationSource><EquationSource Format="MATHML"><math><msup><mrow><mi>R</mi></mrow><mrow><mn>2</mn></mrow></msup></math></EquationSource></InlineEquation> = 0.9951 with RMSE = 0.0695 on ROA regression, highlighting the potential of AI-driven ESG analytics for investment assessment and corporate performance evaluation in developed-market settings.</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

AI-driven ESG analytics for economic resilience: a hybrid ACO-CNRC framework for risk and performance prediction in emerging economies

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年8月4日

查看全文
  • Jiancheng Feng

摘要

Environmental, Social, and Governance (ESG) practices are increasingly used to quantify sustainable corporate behavior and are often linked to long-term financial performance and resilience in public equity markets. This study proposes an AI-based predictive framework to analyze the relationship between ESG-related indicators and firm performance using the Kaggle 400k NYSE Random Investments + Financial Ratios dataset, enriched with additional financial and macroeconomic variables. The data are standardized via Z-score normalization, compressed using Principal Component Analysis (PCA), and refined through Recursive Feature Elimination (RFE) to retain the most informative predictors. For prediction, we develop an Optimized Attentive Ant Colony Convolve Neuro Recurrent Net model (ACO-CNRC) that integrates Ant Colony Optimization for configuration search with CNN/RNN sequential modeling and an attention mechanism for improved representation learning. Experiments on the NYSE-based dataset achieve high predictive performance, reaching 99.74% accuracy on the classification task and, \({R}^{2}\)R2 = 0.9951 with RMSE = 0.0695 on ROA regression, highlighting the potential of AI-driven ESG analytics for investment assessment and corporate performance evaluation in developed-market settings.