Learn R Programming

UBayFS (version 1.0)

A User-Guided Bayesian Framework for Ensemble Feature Selection

Description

The framework proposed in Jenul et al., (2022) , together with an interactive Shiny dashboard. 'UBayFS' is an ensemble feature selection technique embedded in a Bayesian statistical framework. The method combines data and user knowledge, where the first is extracted via data-driven ensemble feature selection. The user can control the feature selection by assigning prior weights to features and penalizing specific feature combinations. 'UBayFS' can be used for common feature selection as well as block feature selection.

Copy Link

Version

Install

install.packages('UBayFS')

Monthly Downloads

165

Version

1.0

License

GPL-3

Maintainer

Anna Jenul

Last Published

March 7th, 2023

Functions in UBayFS (1.0)

train

UBayFS feature selection
sampleInitial

Initial feature set sampling using probabilistic Greedy algorithm
print.UBaymodel

Print a UBayFS model
runInteractive

Run an interactive Shiny app for demonstration
setConstraints

Set constraints in UBaymodel object
is.UBayconstraint

Checks whether a list object implements proper UBayFS user constraints
build.UBaymodel

Build an ensemble for UBayFS
evaluateFS

Evaluate a feature set
is.UBaymodel

Check whether an object is a UBaymodel
buildConstraints

Build a constraint system
buildDecorrConstraints

Build decorrelation constraints
bcw

Breast Cancer Wisconsin dataset
group_admissibility

Admissibility for constraint group
build_train_set

Perform stratified data partition.
build.UBayconstraint

Build a customized constraint for UBayFS
setOptim

Set optimization parameters in a UBaymodel object
posteriorExpectation

Posterior expectation of features
print.UBayconstraint

Prints the `UBayconstraint` object
setWeights

Set weights in UBaymodel object