Estimating High-Dimensional Directed Acyclic Graphs with the PC-Algorithm

Abstract

We consider the PC-algorithm (Spirtes et al., 2000) for estimating the skeleton and equivalence class of a very high-dimensional directed acyclic graph (DAG) with corresponding Gaussian distribution. The PC-algorithm is computationally feasible and often very fast for sparse problems with many nodes (variables), and it has the attractive property to automatically achieve high computational efficiency as a function of sparseness of the true underlying DAG. We prove uniform consistency of the algorithm for very high-dimensional, sparse DAGs where the number of nodes is allowed to quickly grow with sample size n, as fast as O(na) for any 0 a n. We also demonstrate the PC-algorithm for simulated data.

Cite

Text

Kalisch and Bühlmann. "Estimating High-Dimensional Directed Acyclic Graphs with the PC-Algorithm." Journal of Machine Learning Research, 2007.

Markdown

[Kalisch and Bühlmann. "Estimating High-Dimensional Directed Acyclic Graphs with the PC-Algorithm." Journal of Machine Learning Research, 2007.](https://mlanthology.org/jmlr/2007/kalisch2007jmlr-estimating/)

BibTeX

@article{kalisch2007jmlr-estimating,
  title     = {{Estimating High-Dimensional Directed Acyclic Graphs with the PC-Algorithm}},
  author    = {Kalisch, Markus and Bühlmann, Peter},
  journal   = {Journal of Machine Learning Research},
  year      = {2007},
  pages     = {613-636},
  volume    = {8},
  url       = {https://mlanthology.org/jmlr/2007/kalisch2007jmlr-estimating/}
}