Subscribe to DSC Newsletter

How to reduce high concordance (more than 85) in logistic regression model?


I am getting a very high concordance in one of my logistic regression model.

Can anybody explain the effect of it in the model or why it is not recommended of having a very high concordance and what steps to follow to reduce it back to 65-70?

Thanks a lot in advance!



Tags: logistic, regression

Views: 2013

Reply to This

Replies to This Discussion

We've posted at the same time on the same model!! Lol!!!

well, I just wanted to post my thoughts - Concordance is just a consequence of the model.. There's no hard n fast rule that it's only got to be 65-70... If it's higher, it's predictive power is better, since it can identify 1's better than 0's..

But I'd like to hear out what others have to say.. Can we really exercise caution on this statistic? is it possible to control concordance/discordance??! What does it mean to have high/low concordance??
From my experience I don't know why a high concordance would not be recommended. On the contrary - it is a indicator that the model you built is a good one!
I can only guess that someone told you at some point that a high concordance implies an overfit model, which is why validations in their many forms are also important. If the model validates against a holdout sample (especially out of sample and out of time), then there is no reason not to aim for very high concordances.

But if you're deadset on reducing concordance, just toss out the regressor(s) with the smallest p-values.


On Data Science Central

© 2021   TechTarget, Inc.   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service