statsmodels / statsmodels/statsmodels

Diagnostics, Specification tests after discrete and GLM

Open
#1,190 2 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

comp-discrete comp-genmod comp-graphics comp-stats topic-diagnostic type-enh
Dominant language
Python
Stars
11.6k
Forks
3.6k
Avg merge
7h 37m
Merged PRs (30d)
96

Description

(intended as umbrella issue)

We need the diagnostics and specification tests for all models, similar to stats.dignostics, stats.influence_outliers and graphics.regressionplots which is currently mainly (only) for OLS

several textbooks list hypothesis tests, plots and descriptive measures for the results for this, GLM, Logit, Probit

  • regressionplots: seems to apply in a very similar way to discrete models including GLM
  • influence_outliers: several statistics are similar or unchanged, but set is smaller
    mostly based on residuals
  • diagnostics: specification tests

related issues:

???


Starting with diagnostics,
I have some tests for heteroscedasticity after binary models, Logit and Probit, and after Poisson in uncommitted scripts.
I haven't checked yet how much we can reuse current tests like the het_xxx and acorr_xxx

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reading the existing stats.diagnostics, stats.influence_outliers, and graphics.regressionplots implementations for OLS, then review the uncommitted heteroscedasticity tests for binary models, Logit, Probit, and Poisson. Determine which current het_xxx and acorr_xxx tests can be reused; done means an agreed, implemented diagnostics and specification-test scope for the additional models.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
data, testing
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.