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

The aimer package implements the aimer algorithm as described in Ding and McDonald 2017. The code below generates data with 100 rows and 1000 columns, uses cross-validation to select the optimal number of columns, eigenvectors, and threshold for the coefficients, and plots a heatmap of the cross-validation error between the remaining two parameters when one of the three is fixed at the optimal level.

library(aimer)
dat = factorModelSim1(100,1000,5,c(5,1),2,.1,.1)
MyCV = findThresholdSelect(dat$X, dat$Y, seq(2, 50, 2))
plot(MyCV)

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