{"id":"arxiv/2603.08939","url":"https://arxiv.org/abs/2603.08939","tags":["math.ST"],"score":0,"title":"Shape-constrained density estimation with Wasserstein projection","domain":"math.ST","source":"arxiv","authors":["Takeru Matsuda","Ting-Kam Leonard Wong"],"reviews":0,"subject":"math.ST","summary":"Statistical inference based on optimal transport offers a different perspective from that of maximum likelihood, and has increasingly gained attention in recent years. In this paper, we study univariate nonparametric shape-constrained density estimation via projection with respect to the $p$-Wasserstein distance, with a focus on the quadratic case $p = 2$. By considering shape constraints given by displacement convex subsets of the Wasserstein space, Wasserstein projection estimation is a convex optimization problem. We focus on two fundamental examples, namely non-increasing densities on…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-03-09","reactions":{}}