I think it would be both valuable and interesting to put also some historical context, because knowing its history it turns out that in statistics it has lots of regression-related applications. Why is it useful? Because if one day the ML specialist (where the LR is mostly limited to just classification) will join a team of, say, biostatisticians working in clinical trial, where the logistic regression is used almost exclusively for regression (rather than classification) related tasks, the confusion may be strong initially "how come!?". Let's broaden the perspective and show that the logistic regression was initially invented to solve purely regression problems and is used this way nowadays every single day by thousands of researchers and statisticians. And since you mention the GLM already, it's a perfect moment to connect the dots!
Very helpful. Thanks
Good efforts. But it would be amazing if the quiz was interactive.
Very helpful. Kindly keep posting such valuable content. Thank you
I think it would be both valuable and interesting to put also some historical context, because knowing its history it turns out that in statistics it has lots of regression-related applications. Why is it useful? Because if one day the ML specialist (where the LR is mostly limited to just classification) will join a team of, say, biostatisticians working in clinical trial, where the logistic regression is used almost exclusively for regression (rather than classification) related tasks, the confusion may be strong initially "how come!?". Let's broaden the perspective and show that the logistic regression was initially invented to solve purely regression problems and is used this way nowadays every single day by thousands of researchers and statisticians. And since you mention the GLM already, it's a perfect moment to connect the dots!
https://www.linkedin.com/pulse/logistic-regression-has-been-since-its-birth-adrian-olszewski-haygf/
or on Medium: https://medium.com/@r.clin.res/is-logistic-regression-a-regression-46dcce4945dd