dfcorbin/MABsim: Run (contextual) Multi-Armed Bandit Simulations
This package contains functions that allow the user to investigate the performance of a particular non-parametric approach to modelling expected reward functions in the contextual multi-armed bandit (MAB) setting. We use Thompson sampling in order to explore and choose the actions, and we partition the context space in order to better approximate the true expected-reward functions.
Getting started
Browse package contents
Package details |
|
|---|---|
| Author | Douglas Corbin |
| Maintainer | Douglas Corbin <doug.corbin@bristol.ac.uk> |
| License | GPL (>= 2) |
| Version | 0.1.0 |
| Package repository | View on GitHub |
| Installation |
Install the latest version of this package by entering the following in R:
|
