https://github.com/biolab/orange3-conformal

Conformal Prediction library for Orange3
https://github.com/biolab/orange3-conformal

Last synced: about 1 month ago
JSON representation

Conformal Prediction library for Orange3

• Host: GitHub
• URL: https://github.com/biolab/orange3-conformal
• Owner: biolab
• License: gpl-3.0
• Created: 2017-02-03T13:52:33.000Z (almost 8 years ago)
• Default Branch: master
• Last Pushed: 2019-05-24T07:41:10.000Z (over 5 years ago)
• Last Synced: 2024-05-12T16:02:26.453Z (7 months ago)
• Language: Python
• Homepage:
• Size: 3.99 MB
• Stars: 10
• Watchers: 10
• Forks: 7
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE.txt

Awesome Lists containing this project

• awesome-conformal-prediction - Orange3 Conformal Prediction - class-probabilistic-classification using Venn-ABERS (Conformal) prediction](https://github.com/valeman/Multi-class-probabilistic-classification) by Valery Manokhin (Royal Holloway, 2022) (Python)

README

        
Orange3 Conformal Prediction

============================

Conformal Prediction is an add-on for [Orange3](http://orange.biolab.si) data mining software package.

It provides an extensive toolset for conformal prediction.

Installation

------------

To install the add-on, run

    python setup.py install

To register this add-on with Orange, but keep the code in the development directory (do not copy it to 

Python's site-packages directory), run

    python setup.py develop

Usage

-----

The library in the add-on can be used in Python scripts. The add-on does not provide any GUI widgets.

The example below evaluates an inductive conformal predictor at 0.1 significance level

on the Iris dataset (spliting it into a training and testing set in ratio 2:1).

The nonconformity scores used by the conformal predictor are based on the probabilities returned

by a Naive Bayes classifier.

    import Orange

    import orangecontrib.conformal as cp

    tab = Orange.data.Table('iris')

    nc = cp.nonconformity.InverseProbability(Orange.classification.NaiveBayesLearner())

    ic = cp.classification.InductiveClassifier(nc)

    r = cp.evaluation.run(ic, 0.1, cp.evaluation.RandomSampler(tab, 2, 1))

    print(r.accuracy())

Documentation

-------------

Please see doc/Orange-ConformalPrediction.pdf. Documentation in other formats can also be built using Sphinx from the doc directory.

Online documentation is available at [https://orange3-conformal.readthedocs.io](https://orange3-conformal.readthedocs.io).