Crandore Hub

Rstg

STG : Feature Selection using STochastic Gates

'STG' is a method for feature selection in neural network. The procedure is based on probabilistic relaxation of the l0 norm of features, or the count of the number of selected features. The framework simultaneously learns either a nonlinear regression or classification function while selecting a small subset of features. Read more: Yamada et al. (2020) <https://proceedings.mlr.press/v119/yamada20a.html>.

Versions across snapshots

VersionRepositoryFileSize
0.0.1 rolling linux/jammy R-4.5 Rstg_0.0.1.tar.gz 12.9 KiB
0.0.1 rolling linux/noble R-4.5 Rstg_0.0.1.tar.gz 12.8 KiB
0.0.1 rolling source/ R- Rstg_0.0.1.tar.gz 2.9 KiB
0.0.1 latest linux/jammy R-4.5 Rstg_0.0.1.tar.gz 12.9 KiB
0.0.1 latest linux/noble R-4.5 Rstg_0.0.1.tar.gz 12.8 KiB
0.0.1 latest source/ R- Rstg_0.0.1.tar.gz 2.9 KiB
0.0.1 2026-04-26 source/ R- Rstg_0.0.1.tar.gz 2.9 KiB
0.0.1 2026-04-23 source/ R- Rstg_0.0.1.tar.gz 2.9 KiB
0.0.1 2026-04-09 windows/windows R-4.5 Rstg_0.0.1.zip 15.8 KiB
0.0.1 2025-04-20 source/ R- Rstg_0.0.1.tar.gz 2.9 KiB

Dependencies (latest)

Imports