Triad-Constraints-for-Learning-Causal-Structure-of-Latent-Variables

Learning the causal structure of latent variables for linear non-Gaussian data.

Main Function

function [G,Name] = NG2P_Main(X,alpha)

Input:

• X: M*N matrix, where M is the number of variables and N is the sample size.
• alpha: significance level of the independence test.

Output:

• G: connected matrix to represent recovered graph structure (including observed and latent variables).
• Name: the name of variables in G.

See more details in the README.txt.

Test Example

One may use the Test_NG2P_Main.m to test our method.

Notes

Our method relies heavily on independence tests, one can adjust some parameters, like kernel width, in the UInd_KCItest.m of the Package KCI, to ensure the accuracy.

CITATION

If you use this code, please cite the following paper:

Cai, Ruichu, Feng Xie, Clark Glymour, Zhifeng Hao, and Kun Zhang. "Triad Constraints for Learning Causal Structure of Latent Variables." In Advances in Neural Information Processing Systems, pp. 12863-12872. 2019.

If you have problems or questions, do not hesitate to send an email to [email protected].