A Julia implementation of the following paper:
"Learning to Benchmark: Determining Best Achievable Misclassification Error from Training Data," by Morteza Noshad, Li Xu, and Alfred Hero.
The original paper is available at https://arxiv.org/abs/1909.07192
The algorithm approximates the optimal classification error for a problem given a dataset through an ensemble technique.