/**
* @param node The variable node in question.
* @return the error node for the given node.
*/
public Node getErrorNode(Node node) {
if (errorNodes.contains(node)) {
return node;
}
int index = variableNodes.indexOf(node);
if (index == -1) {
throw new NullPointerException(
node
+ " is not a node in this model. Perhaps "
+ "it's another node with the same name.");
}
return errorNodes.get(index);
}
/**
* @return The edge coefficient matrix of the model, a la SemIm. Note that this will normally need
* to be transposed, since [a][b] is the edge coefficient for a-->b, not b-->a. Sorry.
* History. THESE ARE PARAMETERS OF THE MODEL--THE ONLY PARAMETERS.
*/
public TetradMatrix edgeCoef() {
List<Node> variableNodes = getVariableNodes();
TetradMatrix edgeCoef = new TetradMatrix(variableNodes.size(), variableNodes.size());
for (Edge edge : edgeParameters.keySet()) {
if (Edges.isBidirectedEdge(edge)) {
continue;
}
Node a = edge.getNode1();
Node b = edge.getNode2();
int aindex = variableNodes.indexOf(a);
int bindex = variableNodes.indexOf(b);
double coef = edgeParameters.get(edge);
edgeCoef.set(aindex, bindex, coef);
}
return edgeCoef;
}
private void buildIndexing(Graph graph) {
this.hashIndices = new HashMap<Node, Integer>();
for (Node node : graph.getNodes()) {
this.hashIndices.put(node, variables.indexOf(node));
}
}
/**
* Constructs a new standardized SEM IM from the freeParameters in the given SEM IM.
*
* @param im Stop asking me for these things! The given SEM IM!!!
* @param initialization CALCULATE_FROM_SEM if the initial values will be calculated from the
* given SEM IM; INITIALIZE_FROM_DATA if data will be simulated from the given SEM,
* standardized, and estimated.
*/
public StandardizedSemIm(SemIm im, Initialization initialization) {
this.semPm = new SemPm(im.getSemPm());
this.semGraph = new SemGraph(semPm.getGraph());
semGraph.setShowErrorTerms(true);
if (semGraph.existsDirectedCycle()) {
throw new IllegalArgumentException("The cyclic case is not handled.");
}
if (initialization == Initialization.CALCULATE_FROM_SEM) {
// This code calculates the new coefficients directly from the old ones.
edgeParameters = new HashMap<Edge, Double>();
List<Node> nodes = im.getVariableNodes();
TetradMatrix impliedCovar = im.getImplCovar(true);
for (Parameter parameter : im.getSemPm().getParameters()) {
if (parameter.getType() == ParamType.COEF) {
Node a = parameter.getNodeA();
Node b = parameter.getNodeB();
int aindex = nodes.indexOf(a);
int bindex = nodes.indexOf(b);
double vara = impliedCovar.get(aindex, aindex);
double stda = Math.sqrt(vara);
double varb = impliedCovar.get(bindex, bindex);
double stdb = Math.sqrt(varb);
double oldCoef = im.getEdgeCoef(a, b);
double newCoef = (stda / stdb) * oldCoef;
edgeParameters.put(Edges.directedEdge(a, b), newCoef);
} else if (parameter.getType() == ParamType.COVAR) {
Node a = parameter.getNodeA();
Node b = parameter.getNodeB();
Node exoa = semGraph.getExogenous(a);
Node exob = semGraph.getExogenous(b);
double covar = im.getErrCovar(a, b) / Math.sqrt(im.getErrVar(a) * im.getErrVar(b));
edgeParameters.put(Edges.bidirectedEdge(exoa, exob), covar);
}
}
} else {
// This code estimates the new coefficients from simulated data from the old model.
DataSet dataSet = im.simulateData(1000, false);
TetradMatrix _dataSet = dataSet.getDoubleData();
_dataSet = DataUtils.standardizeData(_dataSet);
DataSet dataSetStandardized = ColtDataSet.makeData(dataSet.getVariables(), _dataSet);
SemEstimator estimator = new SemEstimator(dataSetStandardized, im.getSemPm());
SemIm imStandardized = estimator.estimate();
edgeParameters = new HashMap<Edge, Double>();
for (Parameter parameter : imStandardized.getSemPm().getParameters()) {
if (parameter.getType() == ParamType.COEF) {
Node a = parameter.getNodeA();
Node b = parameter.getNodeB();
double coef = imStandardized.getEdgeCoef(a, b);
edgeParameters.put(Edges.directedEdge(a, b), coef);
} else if (parameter.getType() == ParamType.COVAR) {
Node a = parameter.getNodeA();
Node b = parameter.getNodeB();
Node exoa = semGraph.getExogenous(a);
Node exob = semGraph.getExogenous(b);
double covar = -im.getErrCovar(a, b) / Math.sqrt(im.getErrVar(a) * im.getErrVar(b));
edgeParameters.put(Edges.bidirectedEdge(exoa, exob), covar);
}
}
}
this.measuredNodes = Collections.unmodifiableList(semPm.getMeasuredNodes());
}
// Trying to optimize the search for 4-cliques a bit.
private Set<Set<Integer>> findPureClusters2(
List<Integer> _variables, Map<Node, Set<Node>> adjacencies) {
System.out.println("Original variables = " + variables);
Set<Set<Integer>> clusters = new HashSet<Set<Integer>>();
List<Integer> allVariables = new ArrayList<Integer>();
Set<Node> foundVariables = new HashSet<Node>();
for (int i = 0; i < this.variables.size(); i++) allVariables.add(i);
for (int x : _variables) {
Node nodeX = variables.get(x);
if (foundVariables.contains(nodeX)) continue;
List<Node> adjX = new ArrayList<Node>(adjacencies.get(nodeX));
adjX.removeAll(foundVariables);
if (adjX.size() < 3) continue;
for (Node nodeY : adjX) {
if (foundVariables.contains(nodeY)) continue;
List<Node> commonXY = new ArrayList<Node>(adjacencies.get(nodeY));
commonXY.retainAll(adjX);
commonXY.removeAll(foundVariables);
for (Node nodeZ : commonXY) {
if (foundVariables.contains(nodeZ)) continue;
List<Node> commonXZ = new ArrayList<Node>(commonXY);
commonXZ.retainAll(adjacencies.get(nodeZ));
commonXZ.removeAll(foundVariables);
for (Node nodeW : commonXZ) {
if (foundVariables.contains(nodeW)) continue;
if (!adjacencies.get(nodeY).contains(nodeW)) {
continue;
}
int y = variables.indexOf(nodeY);
int w = variables.indexOf(nodeW);
int z = variables.indexOf(nodeZ);
Set<Integer> cluster = quartet(x, y, z, w);
// Note that purity needs to be assessed with respect to all of the variables in order
// to
// remove all latent-measure impurities between pairs of latents.
if (pure(cluster, allVariables)) {
O:
for (int o : _variables) {
if (cluster.contains(o)) continue;
cluster.add(o);
if (!clique(cluster, adjacencies)) {
cluster.remove(o);
continue O;
}
// if (!allVariablesDependent(cluster)) {
// cluster.remove(o);
// continue O;
// }
List<Integer> _cluster = new ArrayList<Integer>(cluster);
ChoiceGenerator gen2 = new ChoiceGenerator(_cluster.size(), 4);
int[] choice2;
int count = 0;
while ((choice2 = gen2.next()) != null) {
int x2 = _cluster.get(choice2[0]);
int y2 = _cluster.get(choice2[1]);
int z2 = _cluster.get(choice2[2]);
int w2 = _cluster.get(choice2[3]);
Set<Integer> quartet = quartet(x2, y2, z2, w2);
// Optimizes for large clusters.
if (quartet.contains(o)) {
if (++count > 2) continue O;
}
if (quartet.contains(o) && !pure(quartet, allVariables)) {
cluster.remove(o);
continue O;
}
}
}
System.out.println(
"Cluster found: " + variablesForIndices(new ArrayList<Integer>(cluster)));
clusters.add(cluster);
foundVariables.addAll(variablesForIndices(new ArrayList<Integer>(cluster)));
}
}
}
}
}
return clusters;
}
// Finds clusters of size 4 or higher.
private Set<Set<Integer>> findPureClusters(
List<Integer> _variables, Map<Node, Set<Node>> adjacencies) {
// System.out.println("Original variables = " + variables);
Set<Set<Integer>> clusters = new HashSet<Set<Integer>>();
List<Integer> allVariables = new ArrayList<Integer>();
for (int i = 0; i < this.variables.size(); i++) allVariables.add(i);
VARIABLES:
while (!_variables.isEmpty()) {
if (_variables.size() < 4) break;
for (int x : _variables) {
Node nodeX = variables.get(x);
List<Node> adjX = new ArrayList<Node>(adjacencies.get(nodeX));
adjX.retainAll(variablesForIndices(new ArrayList<Integer>(_variables)));
for (Node node : new ArrayList<Node>(adjX)) {
if (adjacencies.get(node).size() < 3) {
adjX.remove(node);
}
}
if (adjX.size() < 3) {
continue;
}
ChoiceGenerator gen = new ChoiceGenerator(adjX.size(), 3);
int[] choice;
while ((choice = gen.next()) != null) {
Node nodeY = adjX.get(choice[0]);
Node nodeZ = adjX.get(choice[1]);
Node nodeW = adjX.get(choice[2]);
int y = variables.indexOf(nodeY);
int w = variables.indexOf(nodeW);
int z = variables.indexOf(nodeZ);
Set<Integer> cluster = quartet(x, y, z, w);
if (!clique(cluster, adjacencies)) {
continue;
}
// Note that purity needs to be assessed with respect to all of the variables in order to
// remove all latent-measure impurities between pairs of latents.
if (pure(cluster, allVariables)) {
// Collections.shuffle(_variables);
O:
for (int o : _variables) {
if (cluster.contains(o)) continue;
cluster.add(o);
List<Integer> _cluster = new ArrayList<Integer>(cluster);
if (!clique(cluster, adjacencies)) {
cluster.remove(o);
continue O;
}
// if (!allVariablesDependent(cluster)) {
// cluster.remove(o);
// continue O;
// }
ChoiceGenerator gen2 = new ChoiceGenerator(_cluster.size(), 4);
int[] choice2;
int count = 0;
while ((choice2 = gen2.next()) != null) {
int x2 = _cluster.get(choice2[0]);
int y2 = _cluster.get(choice2[1]);
int z2 = _cluster.get(choice2[2]);
int w2 = _cluster.get(choice2[3]);
Set<Integer> quartet = quartet(x2, y2, z2, w2);
// Optimizes for large clusters.
if (quartet.contains(o)) {
if (++count > 50) continue O;
}
if (quartet.contains(o) && !pure(quartet, allVariables)) {
cluster.remove(o);
continue O;
}
}
}
System.out.println(
"Cluster found: " + variablesForIndices(new ArrayList<Integer>(cluster)));
clusters.add(cluster);
_variables.removeAll(cluster);
continue VARIABLES;
}
}
}
break;
}
return clusters;
}