https://github.com/djo/bandit-with-online-clustering

https://github.com/djo/bandit-with-online-clustering

Last synced: 26 days ago
JSON representation

• Host: GitHub
• URL: https://github.com/djo/bandit-with-online-clustering
• Owner: djo
• License: mit
• Created: 2022-06-03T11:47:15.000Z (over 2 years ago)
• Default Branch: master
• Last Pushed: 2023-02-07T22:28:06.000Z (almost 2 years ago)
• Last Synced: 2023-03-16T22:20:30.307Z (over 1 year ago)
• Language: Jupyter Notebook
• Size: 973 KB
• Stars: 0
• Watchers: 1
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Contextual multi-armed bandit problem with online clustering

![Build](https://github.com/djo/bandit-with-online-clustering/workflows/Python%20application/badge.svg)

The framework for numerical experiments to simulate the contextual multi-armed bandit problem

in the environment with online clustering.

Part of the paper [Multi-armed bandit problem with online clustering as side information](https://wiki.helsinki.fi/download/attachments/406850783/dzhoha-abstract.pdf).

Structure of the project and currently implemented algorithms:

||Files|

|-|-|

|Environments|[Protocol](src/environments/non_stationary_stochastic_environment.py)|

||[Bernoulli MAB](src/environments/bernoulli_bandit.py)|

|Policies|[Protocol](src/policies/policy.py)|

||[Uniform Random](src/policies/uniform_random.py)|

||[Discounted Thompson Sampling (Beta distribution)](src/policies/discounted_beta_thompson_sampling.py)|

|Tests|[Test module](src/test/)|

|Notebooks|[Experiments](notebooks/)|

### Development

```

python3 -m venv env

source env/bin/activate

pip install -r requirements.txt

./pychecks.sh

```

MIT License

Copyright (c) 2022 Andrii Dzhoha