Bankruptcy visualization and prediction using neural networks: A study of U.S. commercial banks
作者：
Highlights：
• We combine multilayer perceptrons and self-organizing maps for bankruptcy prediction.
• We calculate the probability of distress up to three years before bankruptcy occurs.
• We develop a tool to assess bank risk in the short, medium and long term.
• Our model outperforms traditional models of bankruptcy prediction.
• Distressed banks are concentrated in real estate loans and have more provisions.
摘要
•We combine multilayer perceptrons and self-organizing maps for bankruptcy prediction.•We calculate the probability of distress up to three years before bankruptcy occurs.•We develop a tool to assess bank risk in the short, medium and long term.•Our model outperforms traditional models of bankruptcy prediction.•Distressed banks are concentrated in real estate loans and have more provisions.
论文关键词：Bankruptcy prediction,Financial crisis,Multilayer perceptron,Neural networks,Self-organizing maps
论文评审过程：Available online 25 November 2014.
论文官网地址：https://doi.org/10.1016/j.eswa.2014.11.025
