This repository contains the Matlab and R code for the real application in the paper: High-dimensional Varying Index Coefficient Models via Stein’s Identity, Sen Na, Zhuoran Yang, Zhaoran Wang, Mladen Kolar.
The data folder contains the data downloaded from https://datadryad.org/resource/doi:10.5061/dryad.1139fm7, which has been carefully introduced by the paper https://www.nature.com/articles/s41437-018-0105-y.pdf.
The Figures folder contains all output figures.
The script main.m contains Matlab code for recovering sparse parameter matrix. We require CVX package to be installed in advance. Please see http://cvxr.com/ for instruction.
The script preprodata.R contains R code for preprocessing the data. It outputs GenDat.csv which is the final dataset we focus on.
Please see the comments in each script for brief introduction of other files.
To use the code, citation to the paper High-dimensional Varying Index Coefficient Models via Stein’s Identity is required.