Paper Image

Completing covariance matrices using auxiliary data

Published on:

8 February 2024

Primary Category:


Paper Authors:

Joseph Steneman,

Giuseppe Vinci


Key Details

Proposes AuxCov method to complete covariance matrix estimates using auxiliary data

Learns linkage between correlations and auxiliary variables via regression

Has a tuning parameter for blending empirical and predicted covariances

Outperforms other matrix completion methods in simulations

Applied to large-scale neural data analysis

AI generated summary

Completing covariance matrices using auxiliary data

This paper proposes a new method called AuxCov to complete estimates of covariance matrices when data are incomplete. It works by using regression to model the relationship between observed correlations and auxiliary variables like distance between neurons. This model is then used to predict missing correlations. AuxCov has just one tuning parameter that can be chosen empirically. Simulations show AuxCov outperforms existing methods, especially when the auxiliary data are more relevant. It's applied to analyze large neuroscience data sets.

Answers from this paper


No comments yet, be the first to start the conversation...

Sign up to comment on this paper

Sign Up