Testing the mixture model hypothesis via spectral gap

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Boedihardjo, March T., Kileel, Joe, Tombs, Vandy
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914593195950080
author Boedihardjo, March T.
Kileel, Joe
Tombs, Vandy
author_facet Boedihardjo, March T.
Kileel, Joe
Tombs, Vandy
contents In this paper, we study the problem of testing whether or not a given probability measure $μ$ on $\mathbb{R}^{d}$ can be decomposed as a mixture of two probability measures whose second order statistics are significantly different. We call this the problem of testing the mixture model hypothesis. To tackle it, we introduce a new set of computable orthogonal invariants of $μ$, namely, the eigenvalues of the 4th moment operator $T_μ$ associated with the measure. We prove that the largest eigenvalue is always an outlier eigenvalue. Further, we show how the first and second largest eigenvalues of $T_μ$ give nonasymptotic bounds for this problem and give a complete resolution of the asymptotic version of the problem under the $L^{8}$-$L^{2}$ equivalence assumption.
format Preprint
id arxiv_https___arxiv_org_abs_2603_03245
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Testing the mixture model hypothesis via spectral gap
Boedihardjo, March T.
Kileel, Joe
Tombs, Vandy
Probability
Statistics Theory
In this paper, we study the problem of testing whether or not a given probability measure $μ$ on $\mathbb{R}^{d}$ can be decomposed as a mixture of two probability measures whose second order statistics are significantly different. We call this the problem of testing the mixture model hypothesis. To tackle it, we introduce a new set of computable orthogonal invariants of $μ$, namely, the eigenvalues of the 4th moment operator $T_μ$ associated with the measure. We prove that the largest eigenvalue is always an outlier eigenvalue. Further, we show how the first and second largest eigenvalues of $T_μ$ give nonasymptotic bounds for this problem and give a complete resolution of the asymptotic version of the problem under the $L^{8}$-$L^{2}$ equivalence assumption.
title Testing the mixture model hypothesis via spectral gap
topic Probability
Statistics Theory
url https://arxiv.org/abs/2603.03245