Skip to main content
Join the official 2019 Python Developers SurveyStart the survey!

A wrapper for easy plots of learning and validation curves

Project description


A Python wrapper built for software engineers and researchers to facilitate easy creation of learning and validation curve plots from scikit-learn.

The module is meant to complement your workflow in scikit-learn and ease the process of evaluating your models.

The module includes many quality of life features that should save you precious time whenever you want to plot a learning curve to check for bias/variance or plot a validation curve to see the effect of tuning a hyperparameter.


For those not familiar with learning curves, check out Andrew Ng’s excellent discussion of their use at

Over the process of writing many research papers and building many models, I found myself using boilerplate code that I would copy paste for almost every project whenever I wanted to plot a learning curve or validation curve to evaluate models.

Hopefully, this module will save you a few minutes each time you need to plot a learning or validation curve so you can focus on other things.


Python’s pip is the recommended method of installation. From the terminal:

$ pip install sk_modelcurves

Example Usage

Generate a learning curve using accuracy as a metric and 5-fold cross validation.

Assumes a sklearn estimator called knn, training data matrix called X and training labels called y:

$ from sk_modelcurves.learning_curve import draw_learning_curve
$ draw_learning_curve(knn, X, y, scoring='accuracy', cv=5)

Generate multiple learning curves for several estimators using F1 score as a metric, 5-fold cross validation, and names for each of the estimators.

Assumes 3 sklearn estimators called knn2, knn20, knn40, training data matrix called X and training labels called y:

$ from sk_modelcurves.learning_curve import draw_learning_curve
$ draw_learning_curve([knn2, knn20, knn40], X, y, scoring='f1', cv=5,
  estimator_titles=['2 Neighbors', '20 Neighbors', '40 Neighbors'])

Many other options are available. Check out the source code docstrings or the upcoming documentation.


sk-modelcurves is tested to work for Python 2.6 and Python 2.7. Python 3.3+ has not been tested and is assumed to not work until tested.

The required dependencies include scikit-learn (of course!), numpy >= 1.6.1, and matplotlib >= 1.1.1.

To run tests, you will need nose >= 1.1.2.


Anyone is welcome!

If you find a bug or would like to discuss a potential feature, please file an issue first.


After installation, you can launch the test suite from outside the source directory (you will need to have the nose package installed):

$ nosetests -v sk_modelcurves

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for sk_modelcurves, version 0.4
Filename, size File type Python version Upload date Hashes
Filename, size sk_modelcurves-0.4-py2-none-any.whl (7.9 kB) File type Wheel Python version py2 Upload date Hashes View hashes
Filename, size sk_modelcurves-0.4.tar.gz (5.6 kB) File type Source Python version None Upload date Hashes View hashes

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN SignalFx SignalFx Supporter DigiCert DigiCert EV certificate StatusPage StatusPage Status page