Maximum Likelihood Estimation v2
Continuation of https://matchmaticians.com/questions/setkfa .
After receive answer I'm still lost, since I'm not sure what pmodel is(see comments section there).
What exactly pmodel is? Why do we work with pmodel, but not pdata, maximizing its parameters, if it's a family(I udnerstand it as an array for example) of all possible distributions, so I'm little confused here.
If you can explain more widely please.
P.S. Question specifically for Mathe

1 Answer
Solution can be found in
https://drive.google.com/file/d/1OU0l9LXxm2dNXtyPipCWbw0Ho69Be4dy/view?usp=drive_link
-
grant access please
-
Anyone with the link can view the file.
-
now yes, previously i couldn't
-
Ok, thx, tipped
-
@Mathe so to sum up, we have some unknown distribution pdata and some set of samples from it, then we choose somehow(I guess based on some basic conclusions) some distribution. And for this distribution based on our samples we will estimate it's parameters by maximizing it's likelihood function, right? For example if I have normal distribution, it's likelihood function I also know, I can't find it's partial derivatives for mean and variance and solve for 0, than paste all samples and that's it?
-
I meant I CAN here "I can't find it's partial derivatives for mean and variance and solve for 0, than paste all samples and that's it? "
-
-
Yes, that is correct. You take partial derivatives and solve for zero.
-
Ok, thanks again for great explanation
-
-
@Mathe Hi, would you like to answer to similar question about Bayesian statistics for the same amount?
-
Yes! Share the question.
-
@Mathe https://matchmaticians.com/questions/qse0ka
-
- 1 Answer
- 429 views
- Pro Bono
I'll give an answer in a minute.
It looks like I can't upload a file, but I can share a link, I believe.
doesnt matter i think