Comparison-based Machine Learning in Python

Comparison-based learning methods are machine learning algorithms using similarity comparisons ("A and B are more similar than C and D") instead of featurized data.

```python from sklearn.datasets import loadiris from sklearn.modelselection import crossvalscore

from cblearn.datasets import makerandomtriplets from cblearn.embedding import SOE

X = loadiris().data triplets = makerandomtriplets(X, resultformat="list-order", size=2000)

estimator = SOE(n_components=2)

Measure the fit with scikit-learn's cross-validation

scores = crossvalscore(estimator, triplets, cv=5) print(f"The 5-fold CV triplet error is {sum(scores) / len(scores)}.")

Estimate the scale on all triplets

embedding = estimator.fit_transform(triplets) print(f"The embedding has shape {embedding.shape}.") ```

Getting Started

• Installation & Quickstart
• Examples.
• User Guide.

Contribute

We are happy about your bug reports, questions or suggestions as Github Issues and code or documentation contributions as Github Pull Requests. Please see our Contributor Guide.

Related packages

There are more Python packages for comparison-based learning:

• metric-learn is a collection of algorithms for metric learning. The weakly supervised algorithms learn from triplets and quadruplets.
• salmon is a package for efficiently collecting triplets in crowd-sourced experiments. The package implements ordinal embedding algorithms and sampling strategies to query the most informative comparisons actively.

Authors and Acknowledgement

cblearn was initiated by current and former members of the Theory of Machine Learning group of Prof. Dr. Ulrike von Luxburg at the University of Tübingen. The leading developer is David-Elias Künstle.

We want to thank all the contributors here on GitHub. This work has been supported by the Machine Learning Cluster of Excellence, funded by EXC number 2064/1 – Project number 390727645. The authors would like to thank the International Max Planck Research School for Intelligent Systems (IMPRS-IS) for supporting David-Elias Künstle.

License

This library is free to use, share, and adapt under the MIT License conditions.

Citation

Please cite our JOSS paper if you publish work using cblearn:

Künstle et al., (2024). cblearn: Comparison-based Machine Learning in Python. Journal of Open Source Software, 9(98), 6139, https://doi.org/10.21105/joss.06139

@article{Künstle2024, doi = {10.21105/joss.06139}, url = {https://doi.org/10.21105/joss.06139}, year = {2024}, publisher = {The Open Journal}, volume = {9}, number = {98}, pages = {6139}, author = {David-Elias Künstle and Ulrike von Luxburg}, title = {cblearn: Comparison-based Machine Learning in Python}, journal = {Journal of Open Source Software} }