csinva/mdl-complexity

MDL Complexity computations and experiments from the paper "Revisiting complexity and the bias-variance tradeoff".

23
/ 100
Experimental

This tool helps researchers and data scientists evaluate the generalization ability of their predictive models, especially when working with limited training data. It takes your training dataset and the model you've built, then computes a 'Minimum Description Length Complexity' score. This score helps you choose the best model by indicating how well it will perform on new, unseen data, often as effectively as cross-validation.

No commits in the last 6 months.

Use this if you need a theoretically sound way to select the best predictive model for your task, especially when you have limited data and want to understand its potential to generalize.

Not ideal if you are looking for a simple, off-the-shelf hyperparameter tuning library without needing to understand underlying model complexity principles.

predictive-modeling model-selection generalization-assessment statistical-learning machine-learning-research
No License Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 6 / 25
Maturity 8 / 25
Community 9 / 25

How are scores calculated?

Stars

18

Forks

2

Language

Jupyter Notebook

License

Last pushed

Jun 12, 2023

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/csinva/mdl-complexity"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.