References & Citations
Statistics > Methodology
Title: Dirichlet process mixtures under affine transformations of the data
(Submitted on 7 Sep 2018 (v1), last revised 2 Jul 2020 (this version, v3))
Abstract: Location-scale Dirichlet process mixtures of Gaussians (DPM-G) have proved extremely useful in dealing with density estimation and clustering problems in a wide range of domains. Motivated by an astronomical application, in this work we address the robustness of DPM-G models to affine transformations of the data, a natural requirement for any sensible statistical method for density estimation and clustering. First, we devise a coherent prior specification of the model which makes posterior inference invariant with respect to affine transformations of the data. Second, we formalise the notion of asymptotic robustness under data transformation and show that mild assumptions on the true data generating process are sufficient to ensure that DPM-G models feature such a property. Our investigation is supported by an extensive simulation study and illustrated by the analysis of an astronomical dataset consisting of physical measurements of stars in the field of the globular cluster NGC 2419.
Submission history
From: Bernardo Nipoti [view email][v1] Fri, 7 Sep 2018 13:30:06 GMT (841kb,D)
[v2] Thu, 14 Nov 2019 16:14:50 GMT (978kb,D)
[v3] Thu, 2 Jul 2020 13:09:31 GMT (1030kb,D)
Link back to: arXiv, form interface, contact.