A short note on learning discrete distributions
Abstract
The goal of this short note is to provide simple proofs for the "folklore facts" on the sample complexity of learning a discrete probability distribution over a known domain of size k to various distances , with error probability δ.
0