R package facilitating the simulation and evaluation of both contextual and context-free Multi-Armed Bandit policies.
Clone or download

README.md

Contextual: Multi-Armed Bandits in R

lifecycle AppVeyor Build Status Build Status codecovLicense: GPL v3 DOI CRAN_Status_Badge

Overview

R package facilitating the simulation and evaluation of context-free and contextual Multi-Armed Bandit policies.

The package has been developed to:

  • Introduce a wider audience to contextual bandit policies' advanced sequential decision strategies.
  • Ease the implementation, evaluation and dissemination of both existing and new contextual Multi-Armed Bandit policies.

Installation

To install R6 from CRAN:

install.packages('contextual')

To install the development version (requires the devtools package):

devtools::install_github('Nth-iteration-labs/contextual')

Documentation

Overview of core classes

Policies and Bandits

Overview of contextual's growing library of contextual and context-free bandit policies:

General Context-free Contextual Other
Random
Oracle





Epsilon-Greedy
Epsilon-First
UCB-tuned
Thompson Sampling
BootstrapTS
Softmax
Gittins
CMAB Naive Epsilon-Greedy
Epoch-Greedy
LinUCB (General, Disjoint, Hybrid)
LinTS
ProbitTS
LogitBTS
GLMUCB
Lock-in Feedback (LiF)






Overview of contextual's bandit library:

Context-free Contextual Offline Continuous
BasicBernoulliBandit
BasicGaussianBandit



ContextualBernoulli
ContextualLogit
ContextualHybrid
ContextualLinear
ContextualWheel
ReplayEvaluator
DirectMethod
PropensityWeighting
DoublyRobust

Continuum




Maintainers

Robin van Emden: author, maintainer* Maurits Kaptein: supervisor*

* Tilburg University / Jheronimus Academy of Data Science.

If you encounter a clear bug, please file a minimal reproducible example on GitHub.