Scholarlink

<p>Currently, listed companies are facing increasing financial risks and urgently need accurate and timely early warnings, but traditional forecasting methods struggle to capture the characteristics of high-dimensional financial data. The backpropagation neural network is used in this paper to pick input features, build the network architecture, and configure the output layer for the early warning job using financial data from listed businesses. In order to achieve global search and model parameter optimization, the conventional backpropagation neural network is then improved using a genetic algorithm. This enhances the model’s ability to fit financial data. The optimized genetic algorithm and back propagation neural network model have achieved remarkable results in the early warning of financial difficulties, improving the early warning accuracy of ST companies by 1.47% and 7.35% for non-ST companies, and the GA-BP prediction model has reached 97.94% accuracy. At the same time, in the training process, the GA-BP model is more efficient than the traditional BP model. 79 training times reaches the preset target error, which is more efficient than the 117 times of the traditional model. The dynamic change of fitness value further confirms the significant advantage of the optimized model in data related indicators. The research provides a reliable financial risk management tool and provides strong support for the sustainable development of enterprises.</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 financial risk early warning model for listed companies based on a GA-BP neural network

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月18日

查看全文
  • Jing Jing

摘要

Currently, listed companies are facing increasing financial risks and urgently need accurate and timely early warnings, but traditional forecasting methods struggle to capture the characteristics of high-dimensional financial data. The backpropagation neural network is used in this paper to pick input features, build the network architecture, and configure the output layer for the early warning job using financial data from listed businesses. In order to achieve global search and model parameter optimization, the conventional backpropagation neural network is then improved using a genetic algorithm. This enhances the model’s ability to fit financial data. The optimized genetic algorithm and back propagation neural network model have achieved remarkable results in the early warning of financial difficulties, improving the early warning accuracy of ST companies by 1.47% and 7.35% for non-ST companies, and the GA-BP prediction model has reached 97.94% accuracy. At the same time, in the training process, the GA-BP model is more efficient than the traditional BP model. 79 training times reaches the preset target error, which is more efficient than the 117 times of the traditional model. The dynamic change of fitness value further confirms the significant advantage of the optimized model in data related indicators. The research provides a reliable financial risk management tool and provides strong support for the sustainable development of enterprises.