public boolean containsParameter(Edge edge) {
if (Edges.isBidirectedEdge(edge)) {
edge =
Edges.bidirectedEdge(
semGraph.getExogenous(edge.getNode1()), semGraph.getExogenous(edge.getNode2()));
}
return edgeParameters.keySet().contains(edge);
}
private int getDistance(Vertex node, Vertex target) {
for (Edges edge : edges) {
if (edge.getStart().equals(node) && edge.getDestination().equals(target)) {
return edge.getWeight();
}
}
throw new RuntimeException("Error");
}
public void setParameterValue(Edge edge, double value) {
if (Edges.isDirectedEdge(edge)) {
setEdgeCoefficient(edge.getNode1(), edge.getNode2(), value);
} else if (Edges.isBidirectedEdge(edge)) {
setErrorCovariance(edge.getNode1(), edge.getNode2(), value);
} else {
throw new IllegalArgumentException(
"Only directed and bidirected edges are supported: " + edge);
}
}
private List<Vertex> getNeighbors(Vertex node) {
List<Vertex> neighbors = new ArrayList<Vertex>();
for (Edges edge : edges) {
if (edge.getStart().equals(node) && !isSettled(edge.getDestination())) {
neighbors.add(edge.getDestination());
}
}
return neighbors;
}
public static Graph erdosRenyiGraph(int n, int e) {
List<Node> nodes = new ArrayList<Node>();
for (int i = 0; i < n; i++) nodes.add(new GraphNode("X" + i));
Graph graph = new EdgeListGraph(nodes);
for (int e0 = 0; e0 < e; e0++) {
int i1 = RandomUtil.getInstance().nextInt(n);
int i2 = RandomUtil.getInstance().nextInt(n);
if (i1 == i2) {
e0--;
continue;
}
Edge edge = Edges.undirectedEdge(nodes.get(i1), nodes.get(i2));
if (graph.containsEdge(edge)) {
e0--;
continue;
}
graph.addEdge(edge);
}
return graph;
}
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;
}
/**
* @return Returns the error covariance matrix of the model. i.e. [a][b] is the covariance of E_a
* and E_b, with [a][a] of course being the variance of E_a. THESE ARE NOT PARAMETERS OF THE
* MODEL; THEY ARE CALCULATED. Note that elements of this matrix may be Double.NaN; this
* indicates that these elements cannot be calculated.
*/
private TetradMatrix errCovar(Map<Node, Double> errorVariances) {
List<Node> variableNodes = getVariableNodes();
List<Node> errorNodes = new ArrayList<Node>();
for (Node node : variableNodes) {
errorNodes.add(semGraph.getExogenous(node));
}
TetradMatrix errorCovar = new TetradMatrix(errorVariances.size(), errorVariances.size());
for (int index = 0; index < errorNodes.size(); index++) {
Node error = errorNodes.get(index);
double variance = getErrorVariance(error);
errorCovar.set(index, index, variance);
}
for (int index1 = 0; index1 < errorNodes.size(); index1++) {
for (int index2 = 0; index2 < errorNodes.size(); index2++) {
Node error1 = errorNodes.get(index1);
Node error2 = errorNodes.get(index2);
Edge edge = semGraph.getEdge(error1, error2);
if (edge != null && Edges.isBidirectedEdge(edge)) {
double covariance = getErrorCovariance(error1, error2);
errorCovar.set(index1, index2, covariance);
}
}
}
return errorCovar;
}
public boolean isViolatedBy(Graph graph) {
for (Edge edge : graph.getEdges()) {
if (!edge.isDirected()) {
continue;
}
Node from = Edges.getDirectedEdgeTail(edge);
Node to = Edges.getDirectedEdgeHead(edge);
if (isForbidden(from.getName(), to.getName())) {
return true;
}
}
return false;
}
/**
* Sets the covariance for the a<->b edge to the given covariance, if within range. Otherwise does
* nothing.
*
* @param a a <-> b
* @param b a <-> b
* @param covar The covariance of a <-> b.
* @return true if the coefficent was set (i.e. was within range), false if not.
*/
public boolean setErrorCovariance(Node a, Node b, final double covar) {
Edge edge = Edges.bidirectedEdge(semGraph.getExogenous(a), semGraph.getExogenous(b));
if (edgeParameters.get(edge) == null) {
throw new IllegalArgumentException("Not a covariance parameter in this model: " + edge);
}
if (editingEdge == null || !edge.equals(editingEdge)) {
range = getParameterRange(edge);
editingEdge = edge;
}
if (covar > range.getLow() && covar < range.getHigh()) {
edgeParameters.put(edge, covar);
return true;
} else {
return false;
}
// if (!paramInBounds(edge, coef)) {
// edgeParameters.put(edge, d);
// return false;
// }
//
// edgeParameters.put(edge, coef);
// return true;
// if (!paramInBounds(edge, covar)) {
// edgeParameters.put(edge, d);
// return false;
// }
//
// edgeParameters.put(edge, covar);
// return true;
}
/**
* Sets the coefficient for the a->b edge to the given coefficient, if within range. Otherwise
* does nothing.
*
* @param a a -> b
* @param b a -> b
* @param coef The coefficient of a -> b.
* @return true if the coefficent was set (i.e. was within range), false if not.
*/
public boolean setEdgeCoefficient(Node a, Node b, final double coef) {
Edge edge = Edges.directedEdge(a, b);
if (edgeParameters.get(edge) == null) {
throw new NullPointerException("Not a coefficient parameter in this model: " + edge);
}
if (editingEdge == null || !edge.equals(editingEdge)) {
range = getParameterRange(edge);
editingEdge = edge;
}
if (coef > range.getLow() && coef < range.getHigh()) {
edgeParameters.put(edge, coef);
return true;
}
return false;
// if (!paramInBounds(edge, coef)) {
// edgeParameters.put(edge, d);
// return false;
// }
//
// edgeParameters.put(edge, coef);
// return true;
}
private boolean isUndirected(Graph graph, Node x, Node y) {
List<Edge> edges = graph.getEdges(x, y);
if (edges.size() == 1) {
Edge edge = graph.getEdge(x, y);
return Edges.isUndirectedEdge(edge);
}
return false;
}
/**
* @param a a->b
* @param b a->b
* @return The coefficient for a->b.
*/
public double getErrorCovariance(Node a, Node b) {
Edge edge = Edges.bidirectedEdge(semGraph.getExogenous(a), semGraph.getExogenous(b));
Double d = edgeParameters.get(edge);
if (d == null) {
throw new IllegalArgumentException("Not a covariance parameter in this model: " + edge);
}
return d;
}
/**
* @param a a->b
* @param b a->b
* @return The coefficient for a->b.
*/
public double getEdgeCoefficient(Node a, Node b) {
Edge edge = Edges.directedEdge(a, b);
Double d = edgeParameters.get(edge);
if (d == null) {
return Double.NaN;
// throw new IllegalArgumentException("Not a directed edge in this model: " +
// edge);
}
return d;
}
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;
}
private void ruleR1(Graph skeleton, Graph graph, List<Node> nodes) {
for (Node node : nodes) {
SortedMap<Double, String> scoreReports = new TreeMap<Double, String>();
List<Node> adj = skeleton.getAdjacentNodes(node);
DepthChoiceGenerator gen = new DepthChoiceGenerator(adj.size(), adj.size());
int[] choice;
double maxScore = Double.NEGATIVE_INFINITY;
List<Node> parents = null;
while ((choice = gen.next()) != null) {
List<Node> _parents = GraphUtils.asList(choice, adj);
double score = score(node, _parents);
scoreReports.put(-score, _parents.toString());
if (score > maxScore) {
maxScore = score;
parents = _parents;
}
}
for (double score : scoreReports.keySet()) {
TetradLogger.getInstance()
.log(
"score",
"For " + node + " parents = " + scoreReports.get(score) + " score = " + -score);
}
TetradLogger.getInstance().log("score", "");
if (parents == null) {
continue;
}
if (normal(node, parents)) continue;
for (Node _node : adj) {
if (parents.contains(_node)) {
Edge parentEdge = Edges.directedEdge(_node, node);
if (!graph.containsEdge(parentEdge)) {
graph.addEdge(parentEdge);
}
}
}
}
}
/** @return a string representation of the coefficients and variances of the model. */
public String toString() {
StringBuilder buf = new StringBuilder();
NumberFormat nf = NumberFormatUtil.getInstance().getNumberFormat();
buf.append("\nStandardized SEM:");
buf.append("\n\nEdge coefficients (parameters):\n");
for (Edge edge : edgeParameters.keySet()) {
if (!Edges.isDirectedEdge(edge)) {
continue;
}
buf.append("\n" + edge + " " + nf.format(edgeParameters.get(edge)));
}
buf.append("\n\nError covariances (parameters):\n");
for (Edge edge : edgeParameters.keySet()) {
if (!Edges.isBidirectedEdge(edge)) {
continue;
}
buf.append("\n" + edge + " " + nf.format(edgeParameters.get(edge)));
}
buf.append("\n\nError variances (calculated):\n");
for (Node error : getErrorNodes()) {
double variance = getErrorVariance(error);
buf.append("\n" + error + " " + (Double.isNaN(variance) ? "Undefined" : nf.format(variance)));
}
buf.append("\n");
return buf.toString();
}
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.
}
}
}
}
public static Graph weightedRandomGraph(int n, int e) {
List<Node> nodes = new ArrayList<Node>();
for (int i = 0; i < n; i++) nodes.add(new GraphNode("X" + i));
Graph graph = new EdgeListGraph(nodes);
for (int e0 = 0; e0 < e; e0++) {
int i1 = weightedRandom(nodes, graph);
// int i2 = RandomUtil.getInstance().nextInt(n);
int i2 = weightedRandom(nodes, graph);
if (!(shortestPath(nodes.get(i1), nodes.get(i2), graph) < 9)) {
e0--;
continue;
}
if (i1 == i2) {
e0--;
continue;
}
Edge edge = Edges.undirectedEdge(nodes.get(i1), nodes.get(i2));
if (graph.containsEdge(edge)) {
e0--;
continue;
}
graph.addEdge(edge);
}
for (Edge edge : graph.getEdges()) {
Node n1 = edge.getNode1();
Node n2 = edge.getNode2();
if (!graph.isAncestorOf(n2, n1)) {
graph.removeEdge(edge);
graph.addDirectedEdge(n1, n2);
} else {
graph.removeEdge(edge);
graph.addDirectedEdge(n2, n1);
}
}
return graph;
}
/** 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;
}
/**
* 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;
}
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 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);
}
}
}
/**
* @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;
}
/**
* 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());
}
public ParameterRange getCoefficientRange(Node a, Node b) {
return getParameterRange(Edges.directedEdge(a, b));
}
public ParameterRange getCovarianceRange(Node a, Node b) {
return getParameterRange(
Edges.bidirectedEdge(semGraph.getExogenous(a), semGraph.getExogenous(b)));
}
/**
* @param edge a->b or a<->b.
* @return the range of the covariance parameter for a->b or a<->b.
*/
public ParameterRange getParameterRange(Edge edge) {
if (Edges.isBidirectedEdge(edge)) {
edge =
Edges.bidirectedEdge(
semGraph.getExogenous(edge.getNode1()), semGraph.getExogenous(edge.getNode2()));
}
if (!(edgeParameters.keySet().contains(edge))) {
throw new IllegalArgumentException("Not an edge in this model: " + edge);
}
double initial = edgeParameters.get(edge);
if (initial == Double.NEGATIVE_INFINITY) {
initial = Double.MIN_VALUE;
} else if (initial == Double.POSITIVE_INFINITY) {
initial = Double.MAX_VALUE;
}
double value = initial;
// look upward for a point that fails.
double high = value + 1;
while (paramInBounds(edge, high)) {
high = value + 2 * (high - value);
if (high == Double.POSITIVE_INFINITY) {
break;
}
}
// find the boundary using binary search.
double rangeHigh;
if (high == Double.POSITIVE_INFINITY) {
rangeHigh = high;
} else {
double low = value;
while (high - low > 1e-10) {
double midpoint = (high + low) / 2.0;
if (paramInBounds(edge, midpoint)) {
low = midpoint;
} else {
high = midpoint;
}
}
rangeHigh = (high + low) / 2.0;
}
// look downard for a point that fails.
double low = value - 1;
while (paramInBounds(edge, low)) {
low = value - 2 * (value - low);
if (low == Double.NEGATIVE_INFINITY) {
break;
}
}
double rangeLow;
if (low == Double.NEGATIVE_INFINITY) {
rangeLow = low;
} else {
// find the boundary using binary search.
high = value;
while (high - low > 1e-10) {
double midpoint = (high + low) / 2.0;
if (paramInBounds(edge, midpoint)) {
high = midpoint;
} else {
low = midpoint;
}
}
rangeLow = (high + low) / 2.0;
}
if (Edges.isDirectedEdge(edge)) {
edgeParameters.put(edge, initial);
} else if (Edges.isBidirectedEdge(edge)) {
edgeParameters.put(edge, initial);
}
return new ParameterRange(edge, value, rangeLow, rangeHigh);
}
/** 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));
}
}
}
}
/**
* Calculates the error variance for the given error node, given all of the coefficient values in
* the model.
*
* @param error An error term in the model--i.e. a variable with NodeType.ERROR.
* @return The value of the error variance, or Double.NaN is the value is undefined.
*/
private double calculateErrorVarianceFromParams(Node error) {
error = semGraph.getNode(error.getName());
Node child = semGraph.getChildren(error).get(0);
List<Node> parents = semGraph.getParents(child);
double otherVariance = 0;
for (Node parent : parents) {
if (parent == error) continue;
double coef = getEdgeCoefficient(parent, child);
otherVariance += coef * coef;
}
if (parents.size() >= 2) {
ChoiceGenerator gen = new ChoiceGenerator(parents.size(), 2);
int[] indices;
while ((indices = gen.next()) != null) {
Node node1 = parents.get(indices[0]);
Node node2 = parents.get(indices[1]);
double coef1, coef2;
if (node1.getNodeType() != NodeType.ERROR) {
coef1 = getEdgeCoefficient(node1, child);
} else {
coef1 = 1;
}
if (node2.getNodeType() != NodeType.ERROR) {
coef2 = getEdgeCoefficient(node2, child);
} else {
coef2 = 1;
}
List<List<Node>> treks = GraphUtils.treksIncludingBidirected(semGraph, node1, node2);
double cov = 0.0;
for (List<Node> trek : treks) {
double product = 1.0;
for (int i = 1; i < trek.size(); i++) {
Node _node1 = trek.get(i - 1);
Node _node2 = trek.get(i);
Edge edge = semGraph.getEdge(_node1, _node2);
double factor;
if (Edges.isBidirectedEdge(edge)) {
factor = edgeParameters.get(edge);
} else if (!edgeParameters.containsKey(edge)) {
factor = 1;
} else if (semGraph.isParentOf(_node1, _node2)) {
factor = getEdgeCoefficient(_node1, _node2);
} else {
factor = getEdgeCoefficient(_node2, _node1);
}
product *= factor;
}
cov += product;
}
otherVariance += 2 * coef1 * coef2 * cov;
}
}
return 1.0 - otherVariance <= 0 ? Double.NaN : 1.0 - otherVariance;
}
/** Constructs a new SemPm from the given SemGraph. */
public GeneralizedSemPm(SemGraph graph) {
if (graph == null) {
throw new NullPointerException("Graph must not be null.");
}
// if (graph.existsDirectedCycle()) {
// throw new IllegalArgumentExcneption("Cycles are not supported.");
// }
// Cannot afford to allow error terms on this graph to be shown or hidden from the outside; must
// make a
// hidden copy of it and make sure error terms are shown.
this.graph = new SemGraph(graph);
this.graph.setShowErrorTerms(true);
for (Edge edge : this.graph.getEdges()) {
if (Edges.isBidirectedEdge(edge)) {
throw new IllegalArgumentException(
"The generalized SEM PM cannot currently deal with bidirected " + "edges. Sorry.");
}
}
this.nodes = Collections.unmodifiableList(this.graph.getNodes());
for (Node node : nodes) {
namesToNodes.put(node.getName(), node);
}
this.variableNodes = new ArrayList<>();
this.measuredNodes = new ArrayList<>();
for (Node variable : this.nodes) {
if (variable.getNodeType() == NodeType.MEASURED
|| variable.getNodeType() == NodeType.LATENT) {
variableNodes.add(variable);
}
if (variable.getNodeType() == NodeType.MEASURED) {
measuredNodes.add(variable);
}
}
this.errorNodes = new ArrayList<>();
for (Node variable : this.variableNodes) {
List<Node> parents = this.graph.getParents(variable);
boolean added = false;
for (Node _node : parents) {
if (_node.getNodeType() == NodeType.ERROR) {
errorNodes.add(_node);
added = true;
break;
}
}
if (!added) {
errorNodes.add(null);
}
}
this.referencedParameters = new HashMap<>();
this.referencedNodes = new HashMap<>();
this.nodeExpressions = new HashMap<>();
this.nodeExpressionStrings = new HashMap<>();
this.parameterExpressions = new HashMap<>();
this.parameterExpressionStrings = new HashMap<>();
this.parameterEstimationInitializationExpressions = new HashMap<>();
this.parameterEstimationInitializationExpressionStrings = new HashMap<>();
this.startsWithParametersTemplates = new HashMap<>();
this.startsWithParametersEstimationInitializationTemplates = new HashMap<>();
this.variableNames = new ArrayList<>();
for (Node _node : variableNodes) variableNames.add(_node.getName());
for (Node _node : errorNodes) variableNames.add(_node.getName());
try {
List<Node> variableNodes = getVariableNodes();
for (Node node : variableNodes) {
if (!this.graph.isParameterizable(node)) continue;
if (nodeExpressions.get(node) != null) {
continue;
}
String variablestemplate = getVariablesTemplate();
String formula =
TemplateExpander.getInstance().expandTemplate(variablestemplate, this, node);
setNodeExpression(node, formula);
Set<String> parameters = getReferencedParameters(node);
String parametersTemplate = getParametersTemplate();
for (String parameter : parameters) {
if (parameterExpressions.get(parameter) == null) {
if (parametersTemplate != null) {
setParameterExpression(parameter, parametersTemplate);
} else if (this.graph.isTimeLagModel()) {
String expressionString = "Split(-0.9, -.1, .1, 0.9)";
setParameterExpression(parameter, expressionString);
setParametersTemplate(expressionString);
} else {
String expressionString = "Split(-1.5, -.5, .5, 1.5)";
setParameterExpression(parameter, expressionString);
setParametersTemplate(expressionString);
}
}
}
for (String parameter : parameters) {
if (parameterEstimationInitializationExpressions.get(parameter) == null) {
if (parametersTemplate != null) {
setParameterEstimationInitializationExpression(parameter, parametersTemplate);
} else if (this.graph.isTimeLagModel()) {
String expressionString = "Split(-0.9, -.1, .1, 0.9)";
setParameterEstimationInitializationExpression(parameter, expressionString);
} else {
String expressionString = "Split(-1.5, -.5, .5, 1.5)";
setParameterEstimationInitializationExpression(parameter, expressionString);
}
}
setStartsWithParametersTemplate("s", "Split(-1.5, -.5, .5, 1.5)");
setStartsWithParametersEstimationInitializaationTemplate(
"s", "Split(-1.5, -.5, .5, 1.5)");
}
}
for (Node node : errorNodes) {
if (node == null) continue;
String template = getErrorsTemplate();
String formula = TemplateExpander.getInstance().expandTemplate(template, this, node);
setNodeExpression(node, formula);
Set<String> parameters = getReferencedParameters(node);
setStartsWithParametersTemplate("s", "U(1, 3)");
setStartsWithParametersEstimationInitializaationTemplate("s", "U(1, 3)");
for (String parameter : parameters) {
setParameterExpression(parameter, "U(1, 3)");
}
}
} catch (ParseException e) {
throw new IllegalStateException("Parse error in constructing initial model.", e);
}
}
