$XX^T$ Matrices With Independent Entries
Arup Bose, Priyanka Sen

TL;DR
This paper establishes general conditions for the existence of the limiting spectral distribution (LSD) of sample covariance matrices with independent entries, extending classical results and analyzing new cases involving structured matrices.
Contribution
It provides a unified framework for the LSD of $XX^T$ with independent entries and extends results to structured matrices like circulant and Toeplitz, with new insights into the associated partition structures.
Findings
LSD exists under broad conditions for matrices with independent entries.
Partition structures related to symmetric partitions determine the moments of the LSD.
Generalizes previous results for structured matrices like circulant, Toeplitz, and Hankel.
Abstract
Let be the (unscaled) sample covariance matrix where is a real matrix with independent entries. It is well known that if the entries of are independent and identically distributed (i.i.d.) with enough moments and , then the limiting spectral distribution (LSD) of converges to a Marenko-Pastur law. Several extensions of this result are also known. We prove a general result on the existence of the LSD of in probability or almost surely, and in particular, many of the above results follow as special cases. At the same time several new LSD results also follow from our general result. The moments of the LSD are quite involved but can be described via a set of partitions. Unlike in the i.i.d. entries case, these partitions are not necessarily non-crossing, but are related to the special symmetric partitions…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsRandom Matrices and Applications · Advanced Algebra and Geometry · Advanced Combinatorial Mathematics
