Scholarlink

<p>This article explores the application of machine learning (ML) in forecasting banking and bank insolvency. It delves into the methodologies employed, examining the effectiveness of ML techniques in predicting financial instability within the banking sector. In this study, the Gaussian Process Algorithm (GPC) model is used, along with three different optimizers: the Victoria Amazonica Optimizer (VAO), the African Vulture Optimizer (AVO), and the Dandelion Optimization Algorithm (DOA). The improvement in prediction accuracy is achieved by combining the GPC model with each of the three optimizers. The results show significant differences in the performance of the combined models. The GPAV model presents the best performance with an accuracy of 97.19% for the B target. This result highlights the outstanding predictive powers of AVO, showing that it is capable of predicting whether a bank is going to fail accordingly. The other models, GPDO and GPAV, present excellent accuracy rates of 96.26% and 95.32%, respectively, under the same conditions. These findings bring out the great potential of both the VAO and DOA in making accurate predictions of bank financial health. On the other hand, the worst performance among the investigated setups is that of the GPC model, which gives an accuracy of just 74.76% under identical conditions. This emphasizes the importance of using sophisticated optimization methods to enhance the prediction value of the GPC model in predicting probable bank failures. The work concludes by emphasizing the importance of ML, especially the integration of the GPC model with various optimizers, in predicting bank insolvency. The findings highlight the unique capabilities of various optimization algorithms and the need to select the right mix to achieve optimal forecast accuracy when analyzing banks’ financial health.</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

Machine learning models for predicting financial distress focusing on bankruptcy and non-bankruptcy scenarios

  • 期刊论文

  • 发表日期：

    2025年6月16日

查看全文
  • Yingjian Su,
  • Yuehua Bai

摘要

This article explores the application of machine learning (ML) in forecasting banking and bank insolvency. It delves into the methodologies employed, examining the effectiveness of ML techniques in predicting financial instability within the banking sector. In this study, the Gaussian Process Algorithm (GPC) model is used, along with three different optimizers: the Victoria Amazonica Optimizer (VAO), the African Vulture Optimizer (AVO), and the Dandelion Optimization Algorithm (DOA). The improvement in prediction accuracy is achieved by combining the GPC model with each of the three optimizers. The results show significant differences in the performance of the combined models. The GPAV model presents the best performance with an accuracy of 97.19% for the B target. This result highlights the outstanding predictive powers of AVO, showing that it is capable of predicting whether a bank is going to fail accordingly. The other models, GPDO and GPAV, present excellent accuracy rates of 96.26% and 95.32%, respectively, under the same conditions. These findings bring out the great potential of both the VAO and DOA in making accurate predictions of bank financial health. On the other hand, the worst performance among the investigated setups is that of the GPC model, which gives an accuracy of just 74.76% under identical conditions. This emphasizes the importance of using sophisticated optimization methods to enhance the prediction value of the GPC model in predicting probable bank failures. The work concludes by emphasizing the importance of ML, especially the integration of the GPC model with various optimizers, in predicting bank insolvency. The findings highlight the unique capabilities of various optimization algorithms and the need to select the right mix to achieve optimal forecast accuracy when analyzing banks’ financial health.