Iterative Spectral Method (ISM) was originally proposed in the paper Iterative Spectral Method for Alternative Clustering, Wu et al., AISTATS 2018. It is an optimization technique that could be used to solve many HSIC related problems with a Gaussian Kernel. This version is written as an easy extension to other kernels.
The code is written with python3 on a Mint Linux Machine version 17.1. The following libraries must be included.
- numpy
- matplotlib
- sklearn
- scipy
Just run ./sdr.py to see it in action.
This project is licensed under the MIT License - see the LICENSE.md file for details
https://arxiv.org/abs/1909.03093
If you intend to use our code in your research, please cite our paper as follows:
@inproceedings{wu2017ISM,
title={Iterative Spectral Method for Alternative Clustering},
author={Wu, Chieh and Ioannidis, Stratis and Mario, Sznaier and Xiangyu, Li and David, Kaeli and Jennifer, Dy},
booktitle={Artificial Intelligence and Statistics},
year={2018}
}
and
- Chieh Wu , Jared Miller , Mario Sznaier , Yale Chang , Jennifer Dy
We would like to acknowledge support for this project from the NSF grant IIS-1546428. We would also like to thank Zulqarnain Khan for his insightful discussions.