UCSC-SOE-22-04: Sketching in High Dimensional Regression With Big Data Using Gaussian Scale Mixture Priors

Rajarshi Guhaniyogi and Aaron Wolfe Sceffler
05/28/2022 03:51 PM
Bayesian computation of high dimensional linear regression models with popular Gaussian scale mixture prior distributions using Markov Chain Monte Carlo (MCMC) or its variants can be extremely slow or completely prohibitive due to the heavy computational cost that grows in cubic order of p, with p as the number of predictors. Although a few recently developed algorithms make the computation efficient in presence of a small to moderately large sample size (with the complexity growing in the cubic order of n), the computation becomes intractable when sample size n is also large. In this article we adopt the data sketching approach to compress the n original samples by a random linear transformation to m samples in p dimensions, and compute Bayesian regression with Gaussian scale mixture prior distributions with the randomly compressed response vector and predictor matrix. Our proposed approach yields computational complexity growing in the cubic order of m. Another important motivation for this compression procedure is that it anonymizes the data by revealing little information about the original data in the course of analysis. Our detailed empirical investigation with the Horseshoe prior from the class of Gaussian scale mixture priors shows closely similar inference and a massive reduction in per iteration computation time of the proposed approach compared to the regression with the full sample. One notable contribution of this article is to derive posterior contraction rate for high dimensional predictor coefficient with a general class of shrinkage priors on them under data compression/sketching. In particular, we characterize the dimension of the compressed response vector m as a function of the sample size, number of predictors and sparsity in the regression to guarantee accurate estimation of predictor coefficients asymptotically, even after data compression.