sispca 1.0.0

Supervised independent subspace principal component analysis

Supervised Independent Subspace Principal Component Analysis (sisPCA)

sispca is a Python package designed to learn linear representations capturing variations associated with factors of interest in high-dimensional data. It extends the Principal Component Analysis (PCA) to multiple subspaces and encourage subspace disentanglement by maximizing the Hilbert-Schmidt Independence Criterion (HSIC). The model is implemented in PyTorch and uses the Lightning framework for training. See the documentation for more details.

For more theoretical connections and applications, please refer to our paper Disentangling Interpretable Factors of Variations with Supervised Independent Subspace Principal Component Analysis.

Installation

Via GitHub (latest version):

pip install git+https://github.com/JiayuSuPKU/sispca.git#egg=sispca

Via PyPI (stable version):

pip install sispca

Getting Started

Basic usage:

from sispca import Supervision, SISPCADataset, SISPCA

Tutorials:

• Feature selection using sisPCA on the Breast Cancer Wisconsin dataset.
• Learning unsupervised residual subspace in simulation.
• Learning interpretable infection subspaces in scRNA-seq data using sisPCA.

For additional details, please refer to the documentation.

Citation

If you find sisPCA useful in your research, please consider citing our paper:

@inproceedings{
  su2024disentangling,
  title={Disentangling Interpretable Factors of Variations with Supervised Independent Subspace Principal Component Analysis},
  author={Jiayu Su, David A. Knowles, and Raul Rabadan},
  booktitle={The Thirty-eighth Annual Conference on Neural Information Processing Systems},
  year={2024},
  url={https://openreview.net/forum?id=AFnSMlye5K}
}