A python package for a distance-based classifier which can use several different distance metrics.

Installation

To install DistClassiPy, run the following command:

pip install distclassipy

Usage

Here's a quick example to get you started with DistClassiPy:

import distclassipy as dcpy
from sklearn.datasets import make_classification

X, y = make_classification(
    n_samples=1000,
    n_features=4,
    n_informative=2,
    n_redundant=0,
    random_state=0,
    shuffle=False,
)
# Example usage of DistanceMetricClassifier
clf = dcpy.DistanceMetricClassifier()
clf.fit(X, y)
print(clf.predict([[0, 0, 0, 0]], metric="canberra"))

# Example usage of EnsembleDistanceClassifier
ensemble_clf = dcpy.EnsembleDistanceClassifier(feat_idx=0)
ensemble_clf.fit(X, y)
print(ensemble_clf.predict(X))

Features

• Distance Metric-Based Classification: Utilizes a variety of distance metrics for classification.
• Customizable for Scientific Goals: Allows fine-tuning based on scientific objectives by selecting appropriate distance metrics and features, enhancing both computational efficiency and model performance.
• Interpretable Results: Offers improved interpretability of classification outcomes by directly using distance metrics and feature importance, making it ideal for scientific applications.
• Efficient and Scalable: Demonstrates lower computational requirements compared to traditional methods like Random Forests, making it suitable for large datasets.
• Open Source and Accessible: Available as an open-source Python package on PyPI, encouraging broad application in astronomy and beyond.
• (NEW) Ensemble Distance Classification: Leverages an ensemble approach to use different distance metrics for each quantile, improving classification performance across diverse data distributions.
• (NEW) Expanded Distance Metrics: DistClassiPy now offers 43 built-in distance metrics, an increase from the previous 18. Additionally, users can still define and use custom distance metrics as needed.

Documentation

For more detailed information about the package and its functionalities, please refer to the official documentation.

Contributing

Contributions are welcome! If you have suggestions for improvements or bug fixes, please feel free to open an issue or submit a pull request.

License

DistClassiPy is released under the GNU General Public License v3.0. See the LICENSE file for more details.

Citation

If you use DistClassiPy in your research or project, please consider citing the paper:

Chaini, S., Mahabal, A., Kembhavi, A., & Bianco, F. B. (2024). Light Curve Classification with DistClassiPy: a new distance-based classifier. Astronomy and Computing. https://doi.org/10.1016/j.ascom.2024.100850.

Bibtex

@ARTICLE{2024A&C....4800850C,
       author = {{Chaini}, S. and {Mahabal}, A. and {Kembhavi}, A. and {Bianco}, F.~B.},
        title = "{Light curve classification with DistClassiPy: A new distance-based classifier}",
      journal = {Astronomy and Computing},
     keywords = {Variable stars (1761), Astronomy data analysis (1858), Open source software (1866), Astrostatistics (1882), Classification (1907), Light curve classification (1954), Astrophysics - Instrumentation and Methods for Astrophysics, Astrophysics - Solar and Stellar Astrophysics, Computer Science - Machine Learning},
         year = 2024,
        month = jul,
       volume = {48},
          eid = {100850},
        pages = {100850},
          doi = {10.1016/j.ascom.2024.100850},
archivePrefix = {arXiv},
       eprint = {2403.12120},
 primaryClass = {astro-ph.IM},
       adsurl = {https://ui.adsabs.harvard.edu/abs/2024A&C....4800850C},
      adsnote = {Provided by the SAO/NASA Astrophysics Data System}
}

Authors

Siddharth Chaini, Ashish Mahabal, Ajit Kembhavi and Federica B. Bianco.