Thompson Sampling

This package was developed for the course of DS223 at American University of Armenia by Elina Israyelyan

The package is created to implement Thompson Sampling algorithm.

Features

• Implement Thompson Sampling with Beta Distribution
• Implement Thompson Sampling with Normal Distribution
• Do both dynamic and static visualizations for the distributions' pdf functions.

Usage

import thompson_sampling

model = thompson_sampling.model.NormalDistribution() 

# fitting the model 
model.fit(data) 

# predicting the best reward giving arm
model.predict()

For further examples check the examples/ directory or visit the documentation website.

References

A Tutorial on Thompson Sampling. Available here [Accessed 14 May 2022] .
Introduction to Thompson Sampling. Available here [Accessed 14 May 2022].
Github repo BabyRobot. Available here