README.md

In dajmcdon/aimer: Amplified, Initially Marginal, Eigenvector Regression

aimer: Amplified Initially Marginal Eigenvector Regression

The aimer package implements the aimer algorithm as described in Ding and McDonald 2017. The main goal is to use marginal regression to select a subset of coefficients, use matrix approximation to estimate the principal components, and then extend those estimates to the original predictor space before thresholding.

The package uses fast C++ routines to quickly perform matrix computations and cross validation for selection of tuning parameters.

The package provides functions for estimating the model, choosing tuning parameters, and generating simulated data as well as methods for prediction, plotting, and extraction.

Installation

The easiest way to install is to write

devtools::install_github('dajmcdon/aimer')

at the R command prompt. If you want to also access the vignette (takes some time to build ~1 minute), you can use

devtools::install_github('dajmcdon/aimer', build_vignettes = TRUE)

Alternatively, you can clone the repository.

The working documents folder

This contains ongoing ideas for theoretical justification of the method. It is not necessary (and will be ignored) when the package is installed in R.

dajmcdon/aimer documentation built on May 6, 2019, 1:31 a.m.