.. image:: https://github.com/sdpython/mlinsights/raw/main/_doc/_static/project_ico.png :target: https://github.com/sdpython/mlinsights/

mlinsights: extensions to scikit-learn

.. image:: https://dev.azure.com/xavierdupre3/mlinsights/_apis/build/status%2Fsdpython.mlinsights%20(2)?branchName=main :target: https://dev.azure.com/xavierdupre3/mlinsights/_build/latest?definitionId=16&branchName=main

.. image:: https://badge.fury.io/py/mlinsights.svg :target: http://badge.fury.io/py/mlinsights

.. image:: https://img.shields.io/badge/license-MIT-blue.svg :alt: MIT License :target: http://opensource.org/licenses/MIT

.. image:: https://codecov.io/github/sdpython/mlinsights/coverage.svg?branch=main :target: https://codecov.io/github/sdpython/mlinsights?branch=main

.. image:: http://img.shields.io/github/issues/sdpython/mlinsights.png :alt: GitHub Issues :target: https://github.com/sdpython/mlinsights/issues

.. image:: https://pepy.tech/badge/mlinsights/month :target: https://pepy.tech/project/mlinsights/month :alt: Downloads

.. image:: https://img.shields.io/github/forks/sdpython/mlinsights.svg :target: https://github.com/sdpython/mlinsights/ :alt: Forks

.. image:: https://img.shields.io/github/stars/sdpython/mlinsights.svg :target: https://github.com/sdpython/mlinsights/ :alt: Stars

.. image:: https://img.shields.io/github/repo-size/sdpython/mlinsights :target: https://github.com/sdpython/mlinsights/ :alt: size

mlinsights extends scikit-learn with a couple of new models, transformers, metrics, plotting. It provides new trainers such as QuantileLinearRegression which trains a linear regression with L1 norm non-linear correlation based on decision trees, or QuantileMLPRegressor a modification of scikit-learn's MLPRegressor which trains a multi-layer perceptron with L1 norm. It also explores PredictableTSNE which trains a supervized model to replicate t-SNE results or a PiecewiseRegression which partitions the data before fitting a model on each bucket. PiecewiseTreeRegressor trains a piecewise regressor using a linear regression on each piece. IntervalRegressor produces confidence interval by using bootstrapping. ApproximateNMFPredictor approximates a NMF to produce prediction without retraining.