This play list is starting out quite well. I gotta go back on review the feature selection for linear models section of ISLR though. I don't remember them ranking stepwise methods over the coals. They did mention that best subset selection (that's the one you couldn't recall IIRC) is computationally very expensive and doesn't tend to get used on data sets with more than 10 predictors.
I'm starting to doubt what universities are for nowadays (to give a certificate after completing exams, it seems). I've learned much more about statistics on UA-cam that in my lectures. Thank you for this
Needs some extra discussion about "Chance patterns" (~3.11) and how ascribe the distinction between 'chance' (random, some function of Urand(0,1)) and pattern (some function of a parameter not yet clear to the analyst) and how we smash those two together and then drag them apart, repeatedly.
OMG 🤭 this is so entertaining, and I'm learning in PLAIN ENGLISH, thank you!!!! you are unusual and brilliant!
I'm laughing so much that I have to stop the video to learn!!! (but the funny thing is I'm learning too!)
This play list is starting out quite well. I gotta go back on review the feature selection for linear models section of ISLR though. I don't remember them ranking stepwise methods over the coals. They did mention that best subset selection (that's the one you couldn't recall IIRC) is computationally very expensive and doesn't tend to get used on data sets with more than 10 predictors.
What about ridge regression and LASSO for feature reduction?
I'm starting to doubt what universities are for nowadays (to give a certificate after completing exams, it seems). I've learned much more about statistics on UA-cam that in my lectures.
Thank you for this
Needs some extra discussion about "Chance patterns" (~3.11) and how ascribe the distinction between 'chance' (random, some function of Urand(0,1)) and pattern (some function of a parameter not yet clear to the analyst) and how we smash those two together and then drag them apart, repeatedly.
I do have videos somewhere on here that go into more detail about overfitting.
3:50 "the red line represent ...". Can you change the graph? Its a green line...
I can't change a video once it's posted.