Logistic Regression Introduction with Tutorial in JMP
Вставка
- Опубліковано 19 лип 2024
- If you are at a university other than UCSD and have found this or any of my other videos to be useful, please do me a favor and send me a note at ProfessorParris@gmail.com indicating your university affiliation and which videos you've found useful.
Thank you! - Dr. Julian Parris
----
Lecture on Logistic Regression and Tutorial on Simple Binary Logistic Regression In JMP
Recording from a live lecture 02/16/2011
Datasets available at: tinyurl.com/3buyp7e
Sincerely grateful for you time and effort in putting this together. Effective and intuitive communication of often muddled information.
Outstanding. Best explanation of logistic regression I've heard. I especially appreciated the frequent reminders/examples of what each equation/variable actually means. Excellent tone, too.
Thanks. Great teaching method/skill. Complicated concepts made easier with real life examples and discussion around practical applications. More excitement in the delivery than most teachers.
God bless you for these posts; a very satisfied PhD candidate from Winnipeg :)!!!
Thanks for this video! I've been looking for a good reference to refresh my memory on the basics of logistic regression, interpretations etc., and this is the best one that I've seen so far.
Excellent video! Extremely helpful. It helped me put a lot into perspective! Great job!
Very good, very clear exposition. Well done.
This is great--really helpful! Thank you.
Great video, very helpful and nicely explained, Thank you so much
Thanks!!! another very satisfied PhD student from BC Why can't my teachers explain it so clearly :(
super helpful!!! wish you around when I took stats as an undergrad.
thank you so much from your help, i got more and more important how you explain,
thanks again and again
THANK YOU, GREAT LECTURE
Thank you very much. I really do have a wonderful lecture. It will help me a great deal
Thank you. very clear explanation.
@mrpapparappa You're so welcome!
I found it - if we want to interpret the beta estimates of a nominal effect the same way as a continuous effect (i.e., e^estimate = OR), the "nominal" variable needs to be0 or 1 and changed to be a continuous variable. Otherwise, JMP treats it differently.
Thank you!! :)
At 32:22, when you show how to figure out odds ratio for c unit change (instead of a single unit, or the entire range), how do you calculate the 95% CI using this technique? thanks
Thanks! Obviously with over 20 000 views your explanations are very appreciated.
Any pointers on how to interpret the beta estimate when the independent variable is nominal?
Great!
@jojito29 My pleasure!
Great demo for using JMP. Thank you. Can you tell me the name of the textbook you're using?
Very interesting presentation - thanks. Does JMP recognize standard SAS code? Could you write some PROC LOGISTIC commands in a window or does it all have to be through this drag-drop GUI?
JMP has its own scripting language (JSL), so you would need jsl not sas code to execute a jmp analysis via code. JMP can run SAS code if you have a connection to a SAS server, but whatever you run is processed by the SAS server, not JMP. For the most part JMP is driven very effectivley via the GUI so it's not necessary to rely on JSL unless you're scripting for reproducibility or custom workflows.
This might be a stupid question, but I've got some trouble figuring out where those probabilities used in the odds still come from (the once denoted pi-hat).
Regarding the probability of an outcome to the model. Would you first create a regular Linear probability model, with robust stanards, and use those for the (pi / 1 - pi) formula? (i.e the LPM gives you the pi values, and you'd use those in the logit?
/surrender Wow, where did you learn all this? Great video. You should add your teaching information in the info box. I'd totally take this class.
Can you change the regressor unit from 1 unit to being 10 units
which software do use to get logistic regression?
Try running R Studio
Is this JMP Pro? Or basic JMP
You should do one for R as well. :)