CumulativeDistributionalOPE, docstring, and refactor - Githubissues

hakuhodo-technologies / scope-rl

SCOPE-RL: A python library for offline reinforcement learning, off-policy evaluation, and selection

https://scope-rl.readthedocs.io/en/latest/

Apache License 2.0

106 stars 10 forks source link

CumulativeDistributionalOPE, docstring, and refactor #4

Closed aiueola closed 2 years ago

aiueola commented 2 years ago

Type of change

[x] New feature
[x] Bugfix
[x] Document / Docstring
[ ] Other

Description

What has been changed?
- Implemented CumulativeDistributionalOPE, which estimates the cumulative distribution function of the trajectory-wise reward and some statistics including variance and conditional value at risk.
How the logic works?

References

reference name

Checklist

[ ] pass unit test (or unnecessary)
[ ] no errors on newly made test cases
[ ] no errors on existing test cases
[x] applied black formatter
[x] no errors on flake8
[x] no warnings
[ ] work in progress

Comments

Any other comments.

Describe in details later.