I am using a logistic regression and my biggest problem is the very low correlations between the dependent and independents. The highest correlation with the dependent variable is a spend variable with a boxcox transformation and yield only a correlation of 0.2.
The results of my model are not very impressive and this, I beleive, it is related to the fact that all of the variables have a correlation lower or equal to 0.2 with the dependent,
the area under the curve is 0.684 and the classification table is below. on final note, I did try different transformations on all independents and interactions as well.
Did someone had a similar experience?
I can't recall if this ever happened to me, but low correlations usually mean poor model fit
After my posting, I have added a new variable from the database to my list of predictors and the distribution was bimodal which led me to think that my sample is including two different kind of population, i have corrected this and correlations are now fine
This is a super good example of how important data visualization is to good analysis.
Steve Denham
Don't miss out on SAS Innovate - Register now for the FREE Livestream!
Can't make it to Vegas? No problem! Watch our general sessions LIVE or on-demand starting April 17th. Hear from SAS execs, best-selling author Adam Grant, Hot Ones host Sean Evans, top tech journalist Kara Swisher, AI expert Cassie Kozyrkov, and the mind-blowing dance crew iLuminate! Plus, get access to over 20 breakout sessions.
Learn how use the CAT functions in SAS to join values from multiple variables into a single value.
Find more tutorials on the SAS Users YouTube channel.