7

I am wondering if anyone knows where I can find data to train a classification model for bankruptcy prediction.

Bankruptcy prediction appears to be a popular topic, and there are tons of papers about it (see Google Scholar). I've searched through a large number of these papers, but unfortunately none of them include a link to the data they used to train their models. I've also contacted a few authors, but have not heard back to date.

Ideally, I am hoping to find data that contains accounting-related features (X) which can be used to predict whether a firm will go bankrupt or not (Y). It would also be nice to find a dataset that has been used in a previous study (so that I have a benchmark for my model). Many of the datasets in the literature appear to contain around N=1000 examples with about P=5-60 features. The only dataset that I have found so far is the qualitative bankruptcy dataset on UCI (which is not great since it has been derived from other data, and processed substantially).

  • 1
    You might try to contact the UCI data contributors and ask them about obtaining other data sets, or data that is more "raw". – philshem Apr 30 '14 at 21:30
5

In the U.S., bankruptcy data is found in several places. Depending on which parameters you decide are important for your tool and algorithm, some of the following datasets should be helpful to you:

The challenge with any predictive modeling is not where to find the data, but which data will be true indicators of the item you are trying to predict. In this case, the data above show the companies that have become bankrupt. Your challenge is finding those factors that led to their bankruptcy...a much more difficult problem. In this case, it may be that data such as:

These may point to the underlying causes and complexity in the area you are researching.

  • Thanks so much! I was really hoping to find a dataset that was used for one of the previous studies that I posted. Given that nothing is available, however, I'm going to put together a dataset from the resources that you posted. – Berk U. May 28 '14 at 20:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.