Undersampled_Graph_Estimation

Algorithms for graph estimation and data generation built on top of gunfolds
Built for Causal UAI 2017 paper Learning Dynamic Structure from Undersampled Data

Python 2.7

Packages/software used in the making of the paper and repo:
gunfolds pycausal
libpgm
GOBNILP

Algorithm status:

(9/23/17)

discrete_data_generation:

Working with latest gunfolds and latest libpgm

continuous algorithms:

SVAR: working (uses gunfolds.tools version)

PC: working (uses gunfolds.tools version)

GC: working

pyfgs: working (not running in parallel due to java vm issues)

discrete algorithms:

DPC: working

GOBNILP: working, requires GOBNILP

MLL: not working, excluded from paper due to poor performance

To Do:

1. explore new functions for fast greedy equivalence search now in latest pycausal release

Acknowledgment

This work was, in part, supported by NSF IIS-1318759 grant.