tensorflow / tensorflow/probability
Feature Request: Order Beta Regression Distribution
Nobody has claimed this yet.
- Dominant language
- Jupyter Notebook
- Stars
- 4.4k
- Forks
- 1.1k
- PR merge metrics
- No merged PRs in 30d
Description
https://osf.io/preprints/socarxiv/2sx6y/ presents a new spin on the beta distribution which could be very useful as a tfp distribution. The idea is to learn a latent variable and two ordered cutpoints, where values of the latent variable below the lower cutpoint are 0, above the upper cutpoint are 1, and in between are a beta distribution. The linked paper suggests this has advantages over a zero-one-inflated beta distribution. This seems like it could be very useful for modeling bounded variables with values on the exact bounds, e.g. scales or pixel intensities.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reading the linked paper to understand the proposed beta regression distribution, including the latent variable and ordered cutpoints. Done means defining the TFP distribution's behavior for values below, between, and above the cutpoints, with project-appropriate validation and tests identified during implementation.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- tensorflow
- Domain
- machine-learning
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100