// ===========================SCORING METHODS===================//
public double scoreDag(Graph graph) {
Graph dag = new EdgeListGraphSingleConnections(graph);
buildIndexing(graph);
double score = 0.0;
for (Node y : dag.getNodes()) {
Set<Node> parents = new HashSet<Node>(dag.getParents(y));
int nextIndex = -1;
for (int i = 0; i < getVariables().size(); i++) {
nextIndex = hashIndices.get(variables.get(i));
}
int parentIndices[] = new int[parents.size()];
Iterator<Node> pi = parents.iterator();
int count = 0;
while (pi.hasNext()) {
Node nextParent = pi.next();
parentIndices[count++] = hashIndices.get(nextParent);
}
if (this.isDiscrete()) {
score += localDiscreteScore(nextIndex, parentIndices);
} else {
score += localSemScore(nextIndex, parentIndices);
}
}
return score;
}
private Graph changeLatentNames(Graph full, Clusters measurements, List<String> latentVarList) {
Graph g2 = null;
try {
g2 = (Graph) new MarshalledObject(full).get();
} catch (IOException e) {
e.printStackTrace();
} catch (ClassNotFoundException e) {
e.printStackTrace();
}
for (int i = 0; i < measurements.getNumClusters(); i++) {
List<String> d = measurements.getCluster(i);
String latentName = latentVarList.get(i);
for (Node node : full.getNodes()) {
if (!(node.getNodeType() == NodeType.LATENT)) {
continue;
}
List<Node> _children = full.getChildren(node);
_children.removeAll(ReidentifyVariables.getLatents(full));
List<String> childNames = getNames(_children);
if (new HashSet<String>(childNames).equals(new HashSet<String>(d))) {
g2.getNode(node.getName()).setName(latentName);
}
}
}
return g2;
}
private boolean existsUnblockedSemiDirectedPath(Node from, Node to, List<Node> cond, Graph G) {
Queue<Node> Q = new LinkedList<Node>();
Set<Node> V = new HashSet<Node>();
Q.offer(from);
V.add(from);
while (!Q.isEmpty()) {
Node t = Q.remove();
if (t == to) return true;
for (Node u : G.getAdjacentNodes(t)) {
Edge edge = G.getEdge(t, u);
Node c = Edges.traverseSemiDirected(t, edge);
if (c == null) continue;
if (cond.contains(c)) continue;
if (c == to) return true;
if (!V.contains(c)) {
V.add(c);
Q.offer(c);
}
}
}
return false;
}
private Graph restrictToEmpiricalLatents(Graph mimStructure, Graph mimbuildStructure) {
Graph _mim = new EdgeListGraph(mimStructure);
for (Node node : mimbuildStructure.getNodes()) {
if (!mimbuildStructure.containsNode(node)) {
_mim.removeNode(node);
}
}
return _mim;
}
private Graph structure(Graph mim) {
List<Node> latents = new ArrayList<Node>();
for (Node node : mim.getNodes()) {
if (node.getNodeType() == NodeType.LATENT) {
latents.add(node);
}
}
return mim.subgraph(latents);
}
// Cannot be done if the graph changes.
public void setInitialGraph(Graph initialGraph) {
initialGraph = GraphUtils.replaceNodes(initialGraph, variables);
out.println("Initial graph variables: " + initialGraph.getNodes());
out.println("Data set variables: " + variables);
if (!new HashSet<Node>(initialGraph.getNodes()).equals(new HashSet<Node>(variables))) {
throw new IllegalArgumentException("Variables aren't the same.");
}
this.initialGraph = initialGraph;
}
private void uncorrelationExogenousVariables() {
Graph graph = getWorkbench().getGraph();
Set<Edge> edges = graph.getEdges();
for (Edge edge : edges) {
if (Edges.isBidirectedEdge(edge)) {
try {
graph.removeEdge(edge);
} catch (Exception e) {
// Ignore.
}
}
}
}
/**
* Forward equivalence search.
*
* @param graph The graph in the state prior to the forward equivalence search.
*/
private void fes(Graph graph, List<Node> nodes) {
TetradLogger.getInstance().log("info", "** FORWARD EQUIVALENCE SEARCH");
lookupArrows = new HashMap<OrderedPair, Set<Arrow>>();
initializeArrowsForward(nodes);
while (!sortedArrows.isEmpty()) {
Arrow arrow = sortedArrows.first();
sortedArrows.remove(arrow);
Node x = arrow.getX();
Node y = arrow.getY();
clearArrow(x, y);
if (graph.isAdjacentTo(x, y)) {
continue;
}
if (!validInsert(x, y, arrow.getHOrT(), arrow.getNaYX(), graph)) {
continue;
}
List<Node> t = arrow.getHOrT();
double bump = arrow.getBump();
Set<Edge> edges = graph.getEdges();
insert(x, y, t, graph, bump);
score += bump;
rebuildPattern(graph);
// Try to avoid duplicating scoring calls. First clear out all of the edges that need to be
// changed,
// then change them, checking to see if they're already been changed. I know, roundabout, but
// there's
// a performance boost.
for (Edge edge : graph.getEdges()) {
if (!edges.contains(edge)) {
reevaluateForward(graph, nodes, edge.getNode1(), edge.getNode2());
}
}
storeGraph(graph);
}
}
/** Evaluate the Insert(X, Y, T) operator (Definition 12 from Chickering, 2002). */
private double insertEval(Node x, Node y, List<Node> t, List<Node> naYX, Graph graph) {
Set<Node> set1 = new HashSet<Node>(naYX);
set1.addAll(t);
List<Node> paY = graph.getParents(y);
set1.addAll(paY);
Set<Node> set2 = new HashSet<Node>(set1);
set1.add(x);
return scoreGraphChange(y, set1, set2);
}
/** Returns true iif the given set forms a clique in the given graph. */
private static boolean isClique(List<Node> nodes, Graph graph) {
for (int i = 0; i < nodes.size() - 1; i++) {
for (int j = i + 1; j < nodes.size(); j++) {
if (!graph.isAdjacentTo(nodes.get(i), nodes.get(j))) {
return false;
}
}
}
return true;
}
/**
* Find all nodes that are connected to Y by an undirected edge that are adjacent to X (that is,
* by undirected or directed edge).
*/
private static List<Node> getNaYX(Node x, Node y, Graph graph) {
List<Edge> yEdges = graph.getEdges(y);
List<Node> nayx = new ArrayList<Node>();
for (Edge edge : yEdges) {
if (!Edges.isUndirectedEdge(edge)) {
continue;
}
Node z = edge.getDistalNode(y);
if (!graph.isAdjacentTo(z, x)) {
continue;
}
nayx.add(z);
}
return nayx;
}
/** Get all nodes that are connected to Y by an undirected edge and not adjacent to X. */
private static List<Node> getTNeighbors(Node x, Node y, Graph graph) {
List<Edge> yEdges = graph.getEdges(y);
List<Node> tNeighbors = new ArrayList<Node>();
for (Edge edge : yEdges) {
if (!Edges.isUndirectedEdge(edge)) {
continue;
}
Node z = edge.getDistalNode(y);
if (graph.isAdjacentTo(z, x)) {
continue;
}
tNeighbors.add(z);
}
return tNeighbors;
}
private Graph condense(Graph mimStructure, Graph mimbuildStructure) {
// System.out.println("Uncondensed: " + mimbuildStructure);
Map<Node, Node> substitutions = new HashMap<Node, Node>();
for (Node node : mimbuildStructure.getNodes()) {
for (Node _node : mimStructure.getNodes()) {
if (node.getName().startsWith(_node.getName())) {
substitutions.put(node, _node);
break;
}
substitutions.put(node, node);
}
}
HashSet<Node> nodes = new HashSet<Node>(substitutions.values());
Graph graph = new EdgeListGraph(new ArrayList<Node>(nodes));
for (Edge edge : mimbuildStructure.getEdges()) {
Node node1 = substitutions.get(edge.getNode1());
Node node2 = substitutions.get(edge.getNode2());
if (node1 == node2) continue;
if (graph.isAdjacentTo(node1, node2)) continue;
graph.addEdge(new Edge(node1, node2, edge.getEndpoint1(), edge.getEndpoint2()));
}
// System.out.println("Condensed: " + graph);
return graph;
}
private void correlateExogenousVariables() {
Graph graph = getWorkbench().getGraph();
if (graph instanceof Dag) {
JOptionPane.showMessageDialog(
JOptionUtils.centeringComp(), "Cannot add bidirected edges to DAG's.");
return;
}
List<Node> nodes = graph.getNodes();
List<Node> exoNodes = new LinkedList<Node>();
for (int i = 0; i < nodes.size(); i++) {
Node node = nodes.get(i);
if (graph.isExogenous(node)) {
exoNodes.add(node);
}
}
for (int i = 0; i < exoNodes.size(); i++) {
loop:
for (int j = i + 1; j < exoNodes.size(); j++) {
Node node1 = exoNodes.get(i);
Node node2 = exoNodes.get(j);
List<Edge> edges = graph.getEdges(node1, node2);
for (int k = 0; k < edges.size(); k++) {
Edge edge = edges.get(k);
if (Edges.isBidirectedEdge(edge)) {
continue loop;
}
}
graph.addBidirectedEdge(node1, node2);
}
}
}
private Graph pickDag(Graph graph) {
SearchGraphUtils.basicPattern(graph, false);
addRequiredEdges(graph);
boolean containsUndirected;
do {
containsUndirected = false;
for (Edge edge : graph.getEdges()) {
if (Edges.isUndirectedEdge(edge)) {
containsUndirected = true;
graph.removeEdge(edge);
Edge _edge = Edges.directedEdge(edge.getNode1(), edge.getNode2());
graph.addEdge(_edge);
}
}
meekOrient(graph, getKnowledge());
} while (containsUndirected);
return graph;
}
/**
* Greedy equivalence search: Start from the empty graph, add edges till model is significant.
* Then start deleting edges till a minimum is achieved.
*
* @return the resulting Pattern.
*/
public Graph search() {
Graph graph;
if (initialGraph == null) {
graph = new EdgeListGraphSingleConnections(getVariables());
} else {
graph = new EdgeListGraphSingleConnections(initialGraph);
}
fireGraphChange(graph);
buildIndexing(graph);
addRequiredEdges(graph);
topGraphs.clear();
storeGraph(graph);
List<Node> nodes = graph.getNodes();
long start = System.currentTimeMillis();
score = 0.0;
// Do forward search.
fes(graph, nodes);
// Do backward search.
bes(graph);
long endTime = System.currentTimeMillis();
this.elapsedTime = endTime - start;
this.logger.log("graph", "\nReturning this graph: " + graph);
this.logger.log("info", "Elapsed time = " + (elapsedTime) / 1000. + " s");
this.logger.flush();
return graph;
}
// Can be done concurrently.
private double deleteEval(Node x, Node y, List<Node> h, List<Node> naYX, Graph graph) {
List<Node> paY = graph.getParents(y);
Set<Node> paYMinuxX = new HashSet<Node>(paY);
paYMinuxX.remove(x);
Set<Node> set1 = new HashSet<Node>(naYX);
set1.removeAll(h);
set1.addAll(paYMinuxX);
Set<Node> set2 = new HashSet<Node>(naYX);
set2.removeAll(h);
set2.addAll(paY);
return scoreGraphChange(y, set1, set2);
}
// Invalid if then nodes or graph changes.
private void calculateArrowsBackward(Node x, Node y, Graph graph) {
if (x == y) {
return;
}
if (!graph.isAdjacentTo(x, y)) {
return;
}
if (!knowledgeEmpty()) {
if (!getKnowledge().noEdgeRequired(x.getName(), y.getName())) {
return;
}
}
List<Node> naYX = getNaYX(x, y, graph);
clearArrow(x, y);
List<Node> _naYX = new ArrayList<Node>(naYX);
DepthChoiceGenerator gen = new DepthChoiceGenerator(_naYX.size(), _naYX.size());
int[] choice;
while ((choice = gen.next()) != null) {
List<Node> H = GraphUtils.asList(choice, _naYX);
if (!knowledgeEmpty()) {
if (!validSetByKnowledge(y, H)) {
continue;
}
}
double bump = deleteEval(x, y, H, naYX, graph);
if (bump > 0.0) {
Arrow arrow = new Arrow(bump, x, y, H, naYX);
sortedArrows.add(arrow);
addLookupArrow(x, y, arrow);
}
}
}
private void initializeArrowsBackward(Graph graph) {
sortedArrows.clear();
lookupArrows.clear();
for (Edge edge : graph.getEdges()) {
Node x = edge.getNode1();
Node y = edge.getNode2();
if (!knowledgeEmpty()) {
if (!getKnowledge().noEdgeRequired(x.getName(), y.getName())) {
continue;
}
}
if (Edges.isDirectedEdge(edge)) {
calculateArrowsBackward(x, y, graph);
} else {
calculateArrowsBackward(x, y, graph);
calculateArrowsBackward(y, x, graph);
}
}
}
private void buildIndexing(Graph graph) {
this.hashIndices = new HashMap<Node, Integer>();
for (Node node : graph.getNodes()) {
this.hashIndices.put(node, variables.indexOf(node));
}
}
public void rtest3() {
Node x = new GraphNode("X");
Node y = new GraphNode("Y");
Node z = new GraphNode("Z");
Node w = new GraphNode("W");
List<Node> nodes = new ArrayList<Node>();
nodes.add(x);
nodes.add(y);
nodes.add(z);
nodes.add(w);
Graph g = new EdgeListGraph(nodes);
g.addDirectedEdge(x, y);
g.addDirectedEdge(x, z);
g.addDirectedEdge(y, w);
g.addDirectedEdge(z, w);
Graph maxGraph = null;
double maxPValue = -1.0;
ICovarianceMatrix maxLatentCov = null;
Graph mim = DataGraphUtils.randomMim(g, 8, 0, 0, 0, true);
// Graph mim = DataGraphUtils.randomSingleFactorModel(5, 5, 8, 0, 0, 0);
Graph mimStructure = structure(mim);
SemPm pm = new SemPm(mim);
System.out.println("\n\nTrue graph:");
System.out.println(mimStructure);
SemImInitializationParams params = new SemImInitializationParams();
params.setCoefRange(0.5, 1.5);
SemIm im = new SemIm(pm, params);
int N = 1000;
DataSet data = im.simulateData(N, false);
CovarianceMatrix cov = new CovarianceMatrix(data);
for (int i = 0; i < 1; i++) {
ICovarianceMatrix _cov = DataUtils.reorderColumns(cov);
List<List<Node>> partition;
FindOneFactorClusters fofc = new FindOneFactorClusters(_cov, TestType.TETRAD_WISHART, .001);
fofc.search();
partition = fofc.getClusters();
System.out.println(partition);
List<String> latentVarList = reidentifyVariables(mim, data, partition, 2);
Mimbuild2 mimbuild = new Mimbuild2();
mimbuild.setAlpha(0.001);
// mimbuild.setMinimumSize(5);
// To test knowledge.
// Knowledge knowledge = new Knowledge2();
// knowledge.setEdgeForbidden("L.Y", "L.W", true);
// knowledge.setEdgeRequired("L.Y", "L.Z", true);
// mimbuild.setKnowledge(knowledge);
Graph mimbuildStructure = mimbuild.search(partition, latentVarList, _cov);
double pValue = mimbuild.getpValue();
System.out.println(mimbuildStructure);
System.out.println("P = " + pValue);
System.out.println("Latent Cov = " + mimbuild.getLatentsCov());
if (pValue > maxPValue) {
maxPValue = pValue;
maxGraph = new EdgeListGraph(mimbuildStructure);
maxLatentCov = mimbuild.getLatentsCov();
}
}
System.out.println("\n\nTrue graph:");
System.out.println(mimStructure);
System.out.println("\nBest graph:");
System.out.println(maxGraph);
System.out.println("P = " + maxPValue);
System.out.println("Latent Cov = " + maxLatentCov);
System.out.println();
}
private void addRequiredEdges(Graph graph) {
if (true) return;
if (knowledgeEmpty()) return;
for (Iterator<KnowledgeEdge> it = getKnowledge().requiredEdgesIterator(); it.hasNext(); ) {
KnowledgeEdge next = it.next();
Node nodeA = graph.getNode(next.getFrom());
Node nodeB = graph.getNode(next.getTo());
if (!graph.isAncestorOf(nodeB, nodeA)) {
graph.removeEdges(nodeA, nodeB);
graph.addDirectedEdge(nodeA, nodeB);
TetradLogger.getInstance()
.log("insertedEdges", "Adding edge by knowledge: " + graph.getEdge(nodeA, nodeB));
}
}
for (Edge edge : graph.getEdges()) {
final String A = edge.getNode1().getName();
final String B = edge.getNode2().getName();
if (knowledge.isForbidden(A, B)) {
Node nodeA = edge.getNode1();
Node nodeB = edge.getNode2();
if (nodeA != null
&& nodeB != null
&& graph.isAdjacentTo(nodeA, nodeB)
&& !graph.isChildOf(nodeA, nodeB)) {
if (!graph.isAncestorOf(nodeA, nodeB)) {
graph.removeEdges(nodeA, nodeB);
graph.addDirectedEdge(nodeB, nodeA);
TetradLogger.getInstance()
.log("insertedEdges", "Adding edge by knowledge: " + graph.getEdge(nodeB, nodeA));
}
}
if (!graph.isChildOf(nodeA, nodeB)
&& getKnowledge().isForbidden(nodeA.getName(), nodeB.getName())) {
if (!graph.isAncestorOf(nodeA, nodeB)) {
graph.removeEdges(nodeA, nodeB);
graph.addDirectedEdge(nodeB, nodeA);
TetradLogger.getInstance()
.log("insertedEdges", "Adding edge by knowledge: " + graph.getEdge(nodeB, nodeA));
}
}
} else if (knowledge.isForbidden(B, A)) {
Node nodeA = edge.getNode2();
Node nodeB = edge.getNode1();
if (nodeA != null
&& nodeB != null
&& graph.isAdjacentTo(nodeA, nodeB)
&& !graph.isChildOf(nodeA, nodeB)) {
if (!graph.isAncestorOf(nodeA, nodeB)) {
graph.removeEdges(nodeA, nodeB);
graph.addDirectedEdge(nodeB, nodeA);
TetradLogger.getInstance()
.log("insertedEdges", "Adding edge by knowledge: " + graph.getEdge(nodeB, nodeA));
}
}
if (!graph.isChildOf(nodeA, nodeB)
&& getKnowledge().isForbidden(nodeA.getName(), nodeB.getName())) {
if (!graph.isAncestorOf(nodeA, nodeB)) {
graph.removeEdges(nodeA, nodeB);
graph.addDirectedEdge(nodeB, nodeA);
TetradLogger.getInstance()
.log("insertedEdges", "Adding edge by knowledge: " + graph.getEdge(nodeB, nodeA));
}
}
}
}
}
/** Do an actual deletion (Definition 13 from Chickering, 2002). */
private void delete(Node x, Node y, List<Node> subset, Graph graph, double bump) {
Edge trueEdge = null;
if (trueGraph != null) {
Node _x = trueGraph.getNode(x.getName());
Node _y = trueGraph.getNode(y.getName());
trueEdge = trueGraph.getEdge(_x, _y);
}
if (log && verbose) {
Edge oldEdge = graph.getEdge(x, y);
String label = trueGraph != null && trueEdge != null ? "*" : "";
TetradLogger.getInstance()
.log(
"deletedEdges",
(graph.getNumEdges() - 1)
+ ". DELETE "
+ oldEdge
+ " "
+ subset
+ " ("
+ bump
+ ") "
+ label);
out.println(
(graph.getNumEdges() - 1)
+ ". DELETE "
+ oldEdge
+ " "
+ subset
+ " ("
+ bump
+ ") "
+ label);
} else {
int numEdges = graph.getNumEdges() - 1;
if (numEdges % 50 == 0) out.println(numEdges);
}
graph.removeEdge(x, y);
for (Node h : subset) {
Edge oldEdge = graph.getEdge(y, h);
graph.removeEdge(y, h);
graph.addDirectedEdge(y, h);
if (log) {
TetradLogger.getInstance()
.log("directedEdges", "--- Directing " + oldEdge + " to " + graph.getEdge(y, h));
}
if (verbose) {
out.println("--- Directing " + oldEdge + " to " + graph.getEdge(y, h));
}
if (Edges.isUndirectedEdge(graph.getEdge(x, h))) {
if (!graph.isAdjacentTo(x, h))
throw new IllegalArgumentException("Not adjacent: " + x + ", " + h);
oldEdge = graph.getEdge(x, h);
graph.removeEdge(x, h);
graph.addDirectedEdge(x, h);
if (log) {
TetradLogger.getInstance()
.log("directedEdges", "--- Directing " + oldEdge + " to " + graph.getEdge(x, h));
}
if (verbose) {
out.println("--- Directing " + oldEdge + " to " + graph.getEdge(x, h));
}
}
}
}
// serial.
private void insert(Node x, Node y, List<Node> t, Graph graph, double bump) {
if (graph.isAdjacentTo(x, y)) {
return; // The initial graph may already have put this edge in the graph.
// throw new IllegalArgumentException(x + " and " + y + " are already adjacent in
// the graph.");
}
Edge trueEdge = null;
if (trueGraph != null) {
Node _x = trueGraph.getNode(x.getName());
Node _y = trueGraph.getNode(y.getName());
trueEdge = trueGraph.getEdge(_x, _y);
}
graph.addDirectedEdge(x, y);
if (log) {
String label = trueGraph != null && trueEdge != null ? "*" : "";
TetradLogger.getInstance()
.log(
"insertedEdges",
graph.getNumEdges()
+ ". INSERT "
+ graph.getEdge(x, y)
+ " "
+ t
+ " "
+ bump
+ " "
+ label);
} else {
int numEdges = graph.getNumEdges() - 1;
if (verbose) {
if (numEdges % 50 == 0) out.println(numEdges);
}
}
if (verbose) {
String label = trueGraph != null && trueEdge != null ? "*" : "";
out.println(
graph.getNumEdges()
+ ". INSERT "
+ graph.getEdge(x, y)
+ " "
+ t
+ " "
+ bump
+ " "
+ label);
} else {
int numEdges = graph.getNumEdges() - 1;
if (verbose) {
if (numEdges % 50 == 0) out.println(numEdges);
}
}
for (Node _t : t) {
Edge oldEdge = graph.getEdge(_t, y);
if (oldEdge == null) throw new IllegalArgumentException("Not adjacent: " + _t + ", " + y);
graph.removeEdge(_t, y);
graph.addDirectedEdge(_t, y);
if (log && verbose) {
TetradLogger.getInstance()
.log("directedEdges", "--- Directing " + oldEdge + " to " + graph.getEdge(_t, y));
out.println("--- Directing " + oldEdge + " to " + graph.getEdge(_t, y));
}
}
}