I'm using economic data to forecast banking crises. They occur in about 14% of the total possible samples.

Using basic summary statistics, one feature shows a banking crisis has taken place within 3 years of hitting a threshold 42% of the time (38 times out of 89 times if memory serves).

However, logistic regression only gave three instances where it predicted with greater than 40% probability that a crisis was forthcoming within 3 years. I'm not sure why that would be.

Logistic Regression results: enter image description here

The LR results seem to drastically underprice the the probability.

The feature is debt and when debt is large and growing fast, it suggests money is being spent inefficienty and frivously and will inevitably go bad.

Regardless, any thoughts on why LR would be so low relative to the summary statistics and how to handle this better?

  • 3
    $\begingroup$ Could you supply enough information to help us know how the LR was applied and to determine whether it was applied correctly? If not, we really don't have the information needed to answer your question. $\endgroup$ – whuber Oct 29 '16 at 20:09
  • $\begingroup$ Thank you for the offer, I'm new to the posting. The easiest thing might be to just post a link to the folder with the script and data files are located. Thanks so much for any help link $\endgroup$ – bdorn Oct 31 '16 at 21:26
  • $\begingroup$ If there is a better way (shortened data set and code block maybe?) to share the details needed, just let me know and I'll edit the main post or be sure to do it next time. I didn't want to risk not giving enough detail so thought this was best. Thanks again for any help. Cheers $\endgroup$ – bdorn Oct 31 '16 at 21:44

Your Answer

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

Browse other questions tagged or ask your own question.