Message to get across “Multi-arm bandits offer an approach to testing, especially web testing, that allows explicit optimization and more rapid decision making than the traditional statistical approach to designing experiments” (Peter Bruce 2020)
CONTENTS
Introduction
introduce post and what I’m writing about (an alternative to ab testing that is very applicable to web testing) providing an example for each approach along the way
AB testing
Introduce very briefly and provide a simple example. Be careful to not to spend too much time in this section since its not the main topic and focus.
Multi-Arm-Bandito
introduce algorithm formally why its better in web testing context simple example (using the psilon-greedy algorithm for an A/B test described in teh text)
Thompson’s sampling uses a Bayesian approach:
same concept using a beta prior instead seems better will read about it and provide a breif example
References
Citation
@online{jutha2023,
author = {Rahim Jutha},
title = {How to Enhance {AB} Testing with the {Multi-arm} {Bandit}},
date = {2023-09-05},
url = {https://rjutha.github.io/posts/2023-08-15-Spam/},
langid = {en}
}