Can anybody please explain LL null intuitively or more precisely?
Hi Garrett,
The logit model is:
log(pi/1-pi) = log(odds) = beta0 + beta1*x1 + beta2*x2 + ...
where log(odds)
is the log-odds of an event occurring (success of en event), beta0
is the intercept, and x1, x2, ...
are the predictors.
Now, LL-Null is the log-odds of a model that has no predictors and contains just the intercept:
log(odds) = beta0
Therefore, LL-Null is the log-odds of success of an event, estimated without reference to any predictors.
Hope this helps!
Best,
The 365 Team