This repository contains source code for the harmonic expectation chapter of Peter M. C. Harrison's (2020) PhD thesis.
The following code installs the custom R packages required for these analyses:
install.packages("remotes")
remotes::install_github(c("pmcharrison/hrep",
"pmcharrison/incon",
"pmcharrison/hvr"))
Other packages can be installed as needed using install.packages
.
The code is intended to be run by sourcing the files in src/pop-behav
in ascending numeric sequence. They save their outputs to output
.
Each of these files can be run from a blank workspace,
assuming that the previous files have been run already
and their results saved to output
.
The original code for 04-ideal-regression.R
computed permutation-based
feature importance metrics that were very time-consuming to evaluate.
These have been expedited by default by setting perm_int_reps = 1
in
04-ideal-regression.R
. To compute these properly one would have to increase
that number to e.g. 100.