Multiclass Active Learning Algorithms with Application in Astronomy.

Contributors

Alasdair Tran, Cheng Soon Ong, Jakub Nabaglo, David Wu, Wei Yen Lee

License

This package is distributed under a a 3-clause ("Simplified" or "New") BSD license.

Source

https://github.com/chengsoonong/mclass-sky

Doc

https://mclearn.readthedocs.io/en/latest/

Publications

Combining Active Learning Suggestions by Alasdair Tran, Cheng Soon Ong, and Christian Wolf

Active Learning with Gaussian Processes by Jakub Nabaglo

Photometric Classification with Thompson Sampling by Alasdair Tran

Cutting-Plane Methods with Active Learning by David Wu

This repository contains a collection of projects related to active learning methods with application in astronomy. Click on one of the links below to go to the directory of a particular project.

1. Combining Active Learning Suggestions by Alasdair Tran, Cheng Soon Ong, and Christian Wolf
2. Active Learning with Gaussian Processes for Photometric Redshift Prediction
3. Cutting-plane Methods with Applications in Convex Optimization and Active Learning
4. Photometric Classification with Thompson Sampling

mclearn is a Python package that implement selected multiclass active learning algorithms, with a focus in astronomical data.

The dependencies are Python 3.4, numpy, pandas, matplotlib, seaborn, ephem, scipy, ipython, and scikit-learn. It's best to first install the Anaconda distribution for Python 3, then install mclearn using pip:

pip install mclearn