Scalable Fitting Methods for Multivariate Gaussian Additive Models with Covariate-dependent Covariance Matrices

Vincenzo Gioia, Department of Economic, Business, Mathematical, Statistical Sciences · 2025

Read Scalable Fitting Methods for Multivariate Gaussian Additive Models with Covariate-dependent Covariance Matrices by Vincenzo Gioia, Department of Economic, Business, Mathematical, Statistical Sciences (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-03368.txt.

Description

Vincenzo Gioia, Matteo Fasiolo, Ruggero Bellio, and Simon Wood gave scalable fitting methods in 2025 for multivariate Gaussian additive models whose covariance can change with covariates. Such models let both the mean and the spread structure of several responses flex with conditions. A hard part is keeping the covariance matrix valid, which they handle by modeling an unconstrained form of it, through a Cholesky-style split or a matrix logarithm.

The number of parameters grows with the square of the response count, so naive fitting chokes in moderate dimensions. Their empirical Bayes methods use lean model structures, sparse derivative systems, and block-wise computing to cut time and memory. The catalog creator field names only the first author plus a department; the head byline governs. Readers get fitting machinery that makes a rich model class usable at real data sizes.

Catalog record

CREATORVincenzo Gioia, Department of Economic, Business, Mathematical, Statistical Sciences
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE104 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-03368

Browse the full library · About the Tower of Babel