StreamingBandit: Experimenting with Bandit Policies
A large number of statistical decision problems in the social sciences and beyond can be framed as a (contextual) multi-armed bandit problem. However, it is notoriously hard to develop and evaluate policies that tackle these types of problems, and to use such policies in applied studies. To address...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Foundation for Open Access Statistics
2020-08-01
|
| Series: | Journal of Statistical Software |
| Subjects: | |
| Online Access: | https://www.jstatsoft.org/index.php/jss/article/view/2881 |
| _version_ | 1819091155967541248 |
|---|---|
| author | Jules Kruijswijk Robin van Emden Petri Parvinen Maurits Kaptein |
| author_facet | Jules Kruijswijk Robin van Emden Petri Parvinen Maurits Kaptein |
| author_sort | Jules Kruijswijk |
| collection | DOAJ |
| description | A large number of statistical decision problems in the social sciences and beyond can be framed as a (contextual) multi-armed bandit problem. However, it is notoriously hard to develop and evaluate policies that tackle these types of problems, and to use such policies in applied studies. To address this issue, this paper introduces StreamingBandit, a Python web application for developing and testing bandit policies in field studies. StreamingBandit can sequentially select treatments using (online) policies in real time. Once StreamingBandit is implemented in an applied context, different policies can be tested, altered, nested, and compared. StreamingBandit makes it easy to apply a multitude of bandit policies for sequential allocation in field experiments, and allows for the quick development and re-use of novel policies. In this article, we detail the implementation logic of StreamingBandit and provide several examples of its use. |
| first_indexed | 2024-12-21T22:35:14Z |
| format | Article |
| id | doaj.art-ddd90b00ac304227b7765add166bc7eb |
| institution | Directory Open Access Journal |
| issn | 1548-7660 |
| language | English |
| last_indexed | 2024-12-21T22:35:14Z |
| publishDate | 2020-08-01 |
| publisher | Foundation for Open Access Statistics |
| record_format | Article |
| series | Journal of Statistical Software |
| spelling | doaj.art-ddd90b00ac304227b7765add166bc7eb2022-12-21T18:47:59ZengFoundation for Open Access StatisticsJournal of Statistical Software1548-76602020-08-0194114710.18637/jss.v094.i091370StreamingBandit: Experimenting with Bandit PoliciesJules KruijswijkRobin van EmdenPetri ParvinenMaurits KapteinA large number of statistical decision problems in the social sciences and beyond can be framed as a (contextual) multi-armed bandit problem. However, it is notoriously hard to develop and evaluate policies that tackle these types of problems, and to use such policies in applied studies. To address this issue, this paper introduces StreamingBandit, a Python web application for developing and testing bandit policies in field studies. StreamingBandit can sequentially select treatments using (online) policies in real time. Once StreamingBandit is implemented in an applied context, different policies can be tested, altered, nested, and compared. StreamingBandit makes it easy to apply a multitude of bandit policies for sequential allocation in field experiments, and allows for the quick development and re-use of novel policies. In this article, we detail the implementation logic of StreamingBandit and provide several examples of its use.https://www.jstatsoft.org/index.php/jss/article/view/2881sequential decision-makingmulti-armed banditdata streamssequential experimentationpython |
| spellingShingle | Jules Kruijswijk Robin van Emden Petri Parvinen Maurits Kaptein StreamingBandit: Experimenting with Bandit Policies Journal of Statistical Software sequential decision-making multi-armed bandit data streams sequential experimentation python |
| title | StreamingBandit: Experimenting with Bandit Policies |
| title_full | StreamingBandit: Experimenting with Bandit Policies |
| title_fullStr | StreamingBandit: Experimenting with Bandit Policies |
| title_full_unstemmed | StreamingBandit: Experimenting with Bandit Policies |
| title_short | StreamingBandit: Experimenting with Bandit Policies |
| title_sort | streamingbandit experimenting with bandit policies |
| topic | sequential decision-making multi-armed bandit data streams sequential experimentation python |
| url | https://www.jstatsoft.org/index.php/jss/article/view/2881 |
| work_keys_str_mv | AT juleskruijswijk streamingbanditexperimentingwithbanditpolicies AT robinvanemden streamingbanditexperimentingwithbanditpolicies AT petriparvinen streamingbanditexperimentingwithbanditpolicies AT mauritskaptein streamingbanditexperimentingwithbanditpolicies |