Log Likehood, which direction is better

in Machine Learning in Python / Binary predictors in a logistic regression

Is the log likehood more different than 0 means it's a better model (more negative)? But usually it's more near 0 than the LL null. Isn't a more negative result make it more near the LL null?

0 answers ( 0 marked as helpful)

Submit an answer

related questions

Erica Ko

Resolved: Question about the change of log(y)

Alessandro Imbrìaco

Resolved: Why the X axis of Y vs Log(X) goes from 14.5 to 17.5?

Doaa Osman

Resolved: log-likelihood

Erica A. Luong

If lower coef std error indicates better estimate, is there convention on interpreting std err value

Oskar Ciolka

If you are wealthy you can afford college with better tutors for longer timespan etc.

Moatz elmalawany

Hello Why we standrdized the inputs ? Although we log transformed the targets