matlab-causal-development

Matlab code called Tetrad Library for causal modeling algorithms developed by the University of Pittsburgh/Carnegie Mellon University Center for Causal Discovery.

Requirements

Matlab

JDK 1.8

Installation

1. install JDK 1.8 and set the JAVA environment

2. launch Matlab with jdk 1.8:

   in Unbutu 14.04, input shell with

   env MATLAB_JAVA=[filepath]/jre matlab
   

   (where [filepath] is your jdk 1.8 installation path.)

3. check the running java version in Matlab

   version -java
   

4. before running the project, add the path of Tetrad jar file in Matlab or add it in the Matlab's java path file

   such as

   javaaddpath('/home/cw/Delight/matlab-causal-development/tetrad-gui-6.5.4-launch.jar');
   

​ (where '/home/cw/Delight/matlab-causal-development/tetrad-gui-6.5.4-launch.jar' is the path of ‘tetrad-gui-6.5.4-launch.jar’)

Example

Generate random Graph and data

import edu.cmu.tetrad.sem.*;
import edu.cmu.tetrad.graph.*;
import edu.cmu.tetrad.search.*;

nodes = 10；
numLatentNodes = 5；
maxNumEdges = 10;  
sample = 1000;
ll=false;
trueG = GraphUtils.randomDag(nodes,numLatentNodes,maxNumEdges,4,2,2,false);

bayesPm = SemPm(trueG);
bayesIm = SemIm(bayesPm);
dataSet = bayesIm.simulateData(L,ll);

Apply FCI on continue data

import edu.cmu.tetrad.*
import java.util.*
import java.lang.*
import edu.cmu.tetrad.data.*
import edu.cmu.tetrad.graph.*
import edu.cmu.tetrad.search.*
import edu.cmu.tetrad.util.*
import edu.cmu.tetrad.search.indtest.*

data=load('data/cheese2.txt');
dataSet = dataToTetradData(data);
alpha =0.01;

independenceTest = IndTestFisherZ(dataSet,alpha);
test = IndTestFisherZ(dataSet, alpha);   % type of independent test
search1 = Fci(test);     % Fci class
resultG =search1.search();   % use Fci to output a PAG