Scholarlink

<p>The rise of Artificial Intelligence (AI) in financial markets offers new avenues for enhancing banking operations and risk management strategies. This paper employs the CatBoost algorithm to predict the recovery of Non-Performing Loans (NPLs) within a three-month timeframe. Using a comprehensive dataset from the Turkish banking industry, the model demonstrates high predictive accuracy and identifies key features impacting NPL recovery. The results demonstrate that repayment behavior, credit limits, and principal NPL amounts play pivotal roles in predicting NPL recoveries. This discovery underlines the potential of AI to revolutionize NPL management by providing banks with actionable insights into optimizing their recovery strategies and improving their overall financial health. Beyond prediction, our research underscores the impact of AI-driven NPL management on banks’ asset quality, profitability, and stock valuation, while contributing to broader financial stability. By mapping the intersection of AI and operational decision-making in NPL management, this paper contributes to the literature on AI applications in financial markets. The findings support the strategic integration of machine learning into credit risk frameworks, offering a data-driven approach to navigating the complexities of modern banking.</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 case study on non-performing loan recovery: applying the CatBoost algorithm in banking to enhance asset quality

  • 期刊论文

  • 发表日期：

    2026年3月17日

查看全文
  • İsmail Cem Özgüler,
  • Murathan Saygılı,
  • Melih Kaleci,
  • Uğur Sakarya

摘要

The rise of Artificial Intelligence (AI) in financial markets offers new avenues for enhancing banking operations and risk management strategies. This paper employs the CatBoost algorithm to predict the recovery of Non-Performing Loans (NPLs) within a three-month timeframe. Using a comprehensive dataset from the Turkish banking industry, the model demonstrates high predictive accuracy and identifies key features impacting NPL recovery. The results demonstrate that repayment behavior, credit limits, and principal NPL amounts play pivotal roles in predicting NPL recoveries. This discovery underlines the potential of AI to revolutionize NPL management by providing banks with actionable insights into optimizing their recovery strategies and improving their overall financial health. Beyond prediction, our research underscores the impact of AI-driven NPL management on banks’ asset quality, profitability, and stock valuation, while contributing to broader financial stability. By mapping the intersection of AI and operational decision-making in NPL management, this paper contributes to the literature on AI applications in financial markets. The findings support the strategic integration of machine learning into credit risk frameworks, offering a data-driven approach to navigating the complexities of modern banking.