Hi,
I am looking on the output of a logistic regression learner node and there are quite some numbers which I cannot find a documentation for.
There are numbers for the standard errors for the individual variables where it appears to be intuitively clear what they mean. Then there is the number for the Log-likelihood.
Also, z-score and P are not explained.
It would be interesting to have a documentation on these numbers in order to assess their meaning.
The log-likelihood seems to be the natural logarithm of a (maximum???) likelihood, but does this number tell me anything? What is the purpose? How is it calculated?
Thanks
Soren