/sci/ - Science & Math

>>10714504
Is linear regression really what you want? Do you know your operation complexity to already be linear?
You should probably do a polynomial regression and see how much error you can eliminate which each additional term. Alternatively look at your algorithm and estimate the operation complexity.
Is it possible to predict? Yes. But what are you trying to do?

>>10714569
Lets just talk about countable infinite sets. So any two sets that are a bijection from N is defined as countable.
You know that they have the same cardinality, but you don't understand why one set can have "more" elements than the other.
The best way I can describe your example is to use indeterminate form. We're removing an infinite number of elements from an infinite set, and (inf - inf) is indeterminate. This can return a set of three different sizes, empty, finite, or countable. You probably could think of a few examples for each case.

The best way I can think to visualize it is to write down the bijections N->Z and N->(N union Z). You can imagine the "slope" of Z to be 1/2 (because it progresses one value after every 2 natural numbers I guess), and the "slope" of (N union Z) to be 1. So while the limit of the magnitude is infinity, we add 1 value to every two intersections.

Idk if this helps