Is it true that if we want to generate a random something, where each possible choice should have the same probablilty of being generated and can be mapped to a point in a 2D square, we can use this method? I.e., generate a uniformly distributed random point in the 2D square, and map it back to the thing we want to generate originally?
aravind00r
I mean that's sorta like the change of basis or sampling function inversion method right. We take our original uniform distribution and stretch it out into some arbitrary space that we are actually working with.
Is it true that if we want to generate a random something, where each possible choice should have the same probablilty of being generated and can be mapped to a point in a 2D square, we can use this method? I.e., generate a uniformly distributed random point in the 2D square, and map it back to the thing we want to generate originally?
I mean that's sorta like the change of basis or sampling function inversion method right. We take our original uniform distribution and stretch it out into some arbitrary space that we are actually working with.