PCA-PMI

Codes for "Part mutual information for quantifying direct associations in networks" https://doi.org/10.1073/pnas.1522586113

"PCA" is nothing to do with Principle Component Analysis. Here, it is PC algorithm (Peter Spirtes, Clark Glymour), a network structure inference algorithm. PMI is part mutual information, a new criteria to estimate condition independence.

How to use the codes

All of the methods we implemented and the methods we want to compare with are in the folder /lib.

Please run make.m to add the path into Matlab Search path temporarily.

If you want to add the path permanently, please run make -p . You can remove the path from "Set Path" bottom.

• cmi.m
  • Computing Conditional Mutual Information with bin method
• pmi.m
  • Computing Part Mutual Information with bin method
• pmiguass.m
  • Computing Part Mutual Information with Gaussian Distribution Approximation
• dcorr.m
  • Computing Distance Correlation
• pdcor
  • Computing Partial Distance Correlation
• graphicalLasso.m
  • Network Structure Inference with graphicalLasso
• pca_pmi.m
  • Network Structure Inference with PC algorithm combining with Part Mutual Information
• kpca_pmi.m
  • Network Structure Inference with PC algorithm combining with Kernelized Part Mutual Information
• pca_cmi.m
  • Network Structure Inference with PC algorithm combining with Conditional Mutual Information
• pcapcc.m
  • Network Structure Inference with PC algorithm combining with Pearson Correlation

Example

Please run configpath.m in the folder example first.

Please read example/README.md to get all the descriptions of codes in example .

Running Environment

• Windows, Unix/Linux, Mac OS
• Matlab (>2013b)

Citation

@article{zhao2016part,
title={Part mutual information for quantifying direct associations in networks},
author={Zhao, Juan and Zhou, Yiwei and Zhang, Xiujun and Chen, Luonan},
journal={Proceedings of the National Academy of Sciences}, volume={113},
number={18},
pages={5130--5135},
year={2016},
publisher={National Acad Sciences}
}

Authors

You can contact either Juan Zhao ([email protected]) or Yiwei Zhou ([email protected]) for any questions about the codes or the paper.

License

Apache License 2.0