I am a novice statistician using R. I am running a BLR to determine the probability of good or bad sites to build a structure. Each location has a composite "score" that was generated from index values created in a previous model. When I run the BLR, my intercept and coefficient seem out of whack. The intercept registers a value of -91.197, and the coefficient is 18.188. Likewise the log-likelihood does not seem right, as it is -2.1503e-07. Therefore when calculating the predicted probability from the Log Units and Raw Units, I am getting extremely low values like 1.024419e-166. Can anyone shed light on what the issue might be? Do I have a bad data set? Am I using the wrong statistical method? Any help is greatly appreciated.
Asked
Active
Viewed 180 times
1
-
1[Separation](http://stats.stackexchange.com/questions/11109/)? – Scortchi - Reinstate Monica May 20 '14 at 18:00