Projection onto the probability simplex: An efficient algorithm with a simple proof, and an application

We provide an elementary proof of a simple, efficient algorithm for computing the Euclidean projection of a point onto the probability simplex. We also show an application in Laplacian K-modes clustering.