Multiscale Graph Correlation

Multiscale Graph Correlation (MGC, pronounced "magic") is a nonparametric approach for independence and k-sample testing. The MGC framework has been extended to include approaches for reliability and reproducibility, including Discriminability.

Currently available in Python, with many other standard independence and k-sample tests included. Development versions in both R (github and CRAN) and MATLAB support only MGC.

Manuscript reproduction for the Discriminability paper can be found at Discriminability Reproduction.

  1. E. W. Bridgeford, S. Wang, Z. Yang, Z. Wang, T. Xu, C. Craddock, J. Dey, G. Kiar, W. Gray-Roncal, C. Coulantoni, C. Douville, C. E. Priebe, B. Caffo, M. Milham, X. Zuo, (CoRR), and J. T. Vogelstein. Big Data Reproducibility: Applications in Brain Imaging and Genomics. bioRxiv, 2020.
  2. R. Mehta, C. Shen, T. Xu, and J. T. Vogelstein. A Consistent Independence Test for Multivariate Time-Series. arxiv, 2019.
  3. Y. Lee, C. Shen, C. E. Priebe, and J. T. Vogelstein. Network dependence testing via diffusion maps and distance-based correlations. Biometrika, 2019.
  4. J. Xiong, C. Shen, J. Arroyo, and J. T. Vogelstein. Graph Independence Testing. arXiv, 2019.
  5. J. T. Vogelstein, E. W. Bridgeford, Q. Wang, C. E. Priebe, M. Maggioni, and C. Shen. Discovering and deciphering relationships across disparate data modalities. eLife, 2019.
  6. C. Shen, C. E. Priebe, and J. T. Vogelstein. From Distance Correlation to Multiscale Graph Correlation. Journal of the American Statistical Association, 2018.
  7. S. Wang, C. Shen, A. Badea, C. E. Priebe, and J. T. Vogelstein. Signal Subgraph Estimation Via Vertex Screening. arXiv, 2018.