Prediction-Focused Mixtures

The dataset used in the paper is a concatenation of M independent mixtures, each with Km components.

BibTex: