I have used Altman's Z-index model to measure the the probability of bankruptcy of indian banks.
Hi Maria-José,
I think the best model for this type of information is in- and outside information of all banks. Without this information your neural network, whether it is the neural network in your brain or that in the RAM of your (com)puter, you can keep on spitting useless numbers until you notice a bank going broke without any apparent reason and without your numbers predicting it..
Foremost, have a look at past bankrupcy's of banks over the globe, and try to find out why they took place. When you can find some patterns in there, you are on your way to a real bankrupcy early warning system. Without a benchmark, a neural network is useless just like any other statistical method. To say it boldly: "Those who don't know about the past, are bound to repeat it".
I think an estimate of the risks taken by the banking bobo's is the most straightforward approach. If you can mold that into some number, I guess a lot of folks will be interested how you developed it.
Cheers,
Frank
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis