Below you’ll find details on software packages that my students and I have built.
sdp: deep nonparametric estimation of discrete conditional distributions via smoothed dyadic partitioning. In a nutshell, this is a superior alternative to the use of Gaussian mixture models as the final output layer of a deep neural network for the purpose of conditional density estimation. By Wesley Tansey.
FDRreg. Tools for false-discovery rate problems, including false-discovery rate regression (whereby covariates can influence
local FDR). An old version is on CRAN, but you’re better off with the latest version of the package from GitHub available using the
install_github command through the
devtools package. See the Read Me for details.
helloPG. An R package skeleton for incorporating the Polya-Gamma distribution into your own code or package. Not a package per se, and thus not on CRAN. This uses the PG simulation routines originally written by Jesse for BayesLogit (see below).