Minimal assumptions for prediction
Thanks for another insightful post. Is "sampling without replacement" part of the assumptions (something we choose)? Or is rather a given as a constraint (something we have to work with)?
As always, it comes back to intentional vs natural randomness.
If you are running some algorithm that needs a sequence of random numbers, you get to pick if you use with or without replacement sampling.
If you are just getting data handed to you by nature, your stuck with the rules nature makes.
I suppose that in science, there can be a bit of an in between through clever experiment design.
> bit of an in between through clever experiment design.
Good point: moving the goal posts may alarm reviewers but is also potentially going to produce new insights.
Thanks for another insightful post. Is "sampling without replacement" part of the assumptions (something we choose)? Or is rather a given as a constraint (something we have to work with)?
As always, it comes back to intentional vs natural randomness.
If you are running some algorithm that needs a sequence of random numbers, you get to pick if you use with or without replacement sampling.
If you are just getting data handed to you by nature, your stuck with the rules nature makes.
I suppose that in science, there can be a bit of an in between through clever experiment design.
> bit of an in between through clever experiment design.
Good point: moving the goal posts may alarm reviewers but is also potentially going to produce new insights.