hkiyomaru/pu-learning

A collection of notebooks that implement algorithms introduced in "Learning from positive and unlabeled data: a survey"

29
/ 100
Experimental

This project provides practical implementations for training machine learning models when you only have a small number of labeled positive examples and a large pool of unlabeled data. It helps you build classifiers to identify rare events or categories, such as disease detection, fraud identification, or specific customer segmentation, when comprehensive negative examples are hard to come by. The output is a trained classifier that can predict whether new, unseen data belongs to the positive class. This is useful for data scientists and machine learning engineers working with imbalanced datasets or limited labeled data.

No commits in the last 6 months.

Use this if you need to build a classification model but only have data for what you're looking for (positive examples) and a lot of 'everything else' (unlabeled data), without clear negative examples.

Not ideal if you have a well-balanced dataset with clearly labeled positive and negative examples, as traditional classification methods would be more straightforward.

imbalanced-classification rare-event-prediction limited-labeled-data data-science machine-learning-research
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 8 / 25
Maturity 16 / 25
Community 5 / 25

How are scores calculated?

Stars

46

Forks

2

Language

Jupyter Notebook

License

MIT

Last pushed

Aug 20, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/hkiyomaru/pu-learning"

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