{"id":"arxiv/2606.29109","url":"https://arxiv.org/abs/2606.29109","tags":["math.ST","stat.ME"],"score":0,"title":"Posterior concentration and adaptation of the mixing measure in Dirichlet process mixtures","domain":"math.ST","source":"arxiv","authors":["Filippo Ascolani"],"reviews":0,"subject":"math.ST","summary":"We study the asymptotic properties of the posterior on the latent space for infinite mixtures driven by a Dirichlet process, both in terms of mixing measure and clustering behaviour. In the well-specified regime, where the data are generated by a finite mixture of location densities, we show that the posterior is adaptive to the true number of components $K$: indeed the cumulative mass assigned to weights of the stick-breaking representation beyond the $K$-th one vanishes as $n^{-1/2}$, up to terms growing slower than any polynomial. This also implies a nearly optimal posterior contraction…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-06-27","reactions":{}}