algo_leverage: algo_leverage function

In vs385/stsci6520: Stsci6520 HW2

Description

The function implements algorithmic leveraging or linear regression using leverage score and uniform-based row subsampling. It replicates the simulation setting of Fig 1 in Ma et Al's in https://onlinelibrary.wiley.com/doi/full/10.1002/wics.1324.

Usage

algo_leverage(y, X, r = 0.1, sample_size = r, method = "leverage")

Arguments

y	The response vector with n entries
X	The covariates n by p matrix
r	Percentage of subsamples to be drwan; default set to 0.1
sample_size	Number of subsamples to consider for regression
method	The type of subsampling: 'unif' for uniform-based, 'leverage' for leverage-score based

Value

The coefficient vector beta

Examples

algo_leverage(rnorm(500), matrix(rnorm(1000),nrow=500), method='unif')

vs385/stsci6520 documentation built on May 21, 2019, 12:35 a.m.