Repo for "Did we personalize? Assessing personalization by an online reinforcement learning algorithm using resampling"

Given access to HeartSteps V2/V3 data,

• Install the environment.yml file.
• Create './output'
• Setup './init' with setup files from 'setup_parameters.py'
• Example Execution of Resampling Script: python main.py -u 0 -b 1 -s 3 -userBIdx 1 -bi Zero where we resample user 0 (-u 0), the second time (-b 1, -userBIdx 1), seed 3 (-s 3), and baseline Zero (-bi Zero).
• Follow Resampling_Paper.ipynb to regenerate the plots in the paper.

Reach out to [email protected] for any questions or additional setup files referenced in the scripts above. Per the IRB and consent agreements for HeartSteps V2/V3, the team is unable to share the HeartSteps V2/V3 data.