/** @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 String reportIfContinuous(Graph dag, DataSet dataSet) {
SemPm semPm = new SemPm(dag);
SemEstimator estimator = new SemEstimator(dataSet, semPm);
estimator.estimate();
SemIm semIm = estimator.getEstimatedSem();
NumberFormat nf = NumberFormat.getInstance();
nf.setMaximumFractionDigits(4);
StringBuilder buf = new StringBuilder();
buf.append("\nDegrees of Freedom = ")
.append(semPm.getDof())
.append("Chi-Square = ")
.append(nf.format(semIm.getChiSquare()))
.append("\nP Value = ")
.append(nf.format(semIm.getPValue()))
.append("\nBIC Score = ")
.append(nf.format(semIm.getBicScore()));
buf.append(
"\n\nThe above chi square test assumes that the maximum "
+ "likelihood function over the measured variables has been "
+ "maximized. Under that assumption, the null hypothesis for "
+ "the test is that the population covariance matrix over all "
+ "of the measured variables is equal to the estimated covariance "
+ "matrix over all of the measured variables written as a function "
+ "of the free model parameters--that is, the unfixed parameters "
+ "for each directed edge (the linear coefficient for that edge), "
+ "each exogenous variable (the variance for the error term for "
+ "that variable), and each bidirected edge (the covariance for "
+ "the exogenous variables it connects). The model is explained "
+ "in Bollen, Structural Equations with Latent Variable, 110. ");
return buf.toString();
}
public String toString() {
StringBuilder buf = new StringBuilder();
buf.append("\n\nRange for " + edge);
buf.append("\nCurrent value = " + coef);
buf.append("\nLow end of range = " + low);
buf.append("\nHigh end of range = " + high);
return buf.toString();
}
/**
* @return a relatively brief String representation of this SEM PM--the equations and
* distributions of the model. Initial value distributions for freeParameters are not printed.
*/
public String toString() {
StringBuilder buf = new StringBuilder();
buf.append("\nEquations:\n");
for (Node node : variableNodes) {
buf.append("\n").append(node).append(" = ").append(nodeExpressionStrings.get(node));
}
buf.append("\n\nErrors:\n");
for (Node node : errorNodes) {
buf.append("\n").append(node).append(" ~ ").append(nodeExpressionStrings.get(node));
}
buf.append("\n\nParameters:\n");
for (String param : getParameters()) {
buf.append("\n").append(param).append(" ~ ").append(getParameterExpressionString(param));
}
return buf.toString();
}
private String reportIfDiscrete(Graph dag, DataSet dataSet) {
List vars = dataSet.getVariables();
Map<String, DiscreteVariable> nodesToVars = new HashMap<String, DiscreteVariable>();
for (int i = 0; i < dataSet.getNumColumns(); i++) {
DiscreteVariable var = (DiscreteVariable) vars.get(i);
String name = var.getName();
Node node = new GraphNode(name);
nodesToVars.put(node.getName(), var);
}
BayesPm bayesPm = new BayesPm(new Dag(dag));
List<Node> nodes = bayesPm.getDag().getNodes();
for (Node node : nodes) {
Node var = nodesToVars.get(node.getName());
if (var instanceof DiscreteVariable) {
DiscreteVariable var2 = nodesToVars.get(node.getName());
int numCategories = var2.getNumCategories();
List<String> categories = new ArrayList<String>();
for (int j = 0; j < numCategories; j++) {
categories.add(var2.getCategory(j));
}
bayesPm.setCategories(node, categories);
}
}
BayesProperties properties = new BayesProperties(dataSet, dag);
properties.setGraph(dag);
NumberFormat nf = NumberFormat.getInstance();
nf.setMaximumFractionDigits(4);
StringBuilder buf = new StringBuilder();
buf.append("\nP-value = ").append(properties.getLikelihoodRatioP());
buf.append("\nDf = ").append(properties.getPValueDf());
buf.append("\nChi square = ").append(nf.format(properties.getPValueChisq()));
buf.append("\nBIC score = ").append(nf.format(properties.getBic()));
buf.append("\n\nH0: Completely disconnected graph.");
return buf.toString();
}
private void saveKnowledge(Writer out) throws IOException {
StringBuilder buf = new StringBuilder();
buf.append("/knowledge");
buf.append("\naddtemporal\n");
for (int i = 0; i < tierSpecs.size(); i++) {
String forbiddenWithin = isTierForbiddenWithin(i) ? "*" : "";
buf.append("\n").append(i).append(forbiddenWithin).append(" ");
List<String> tier = getTier(i);
for (Object aTier : tier) {
String name = (String) aTier;
buf.append(name).append(" ");
}
}
buf.append("\n");
buf.append("\nforbiddirect\n\n");
Set<OrderedPair<Set<MyNode>>> copy = new HashSet<>(forbiddenRulesSpecs);
copy.removeAll(forbiddenTierRules());
for (OrderedPair<Set<MyNode>> o : copy) {
Set<MyNode> first = o.getFirst();
Set<MyNode> second = o.getSecond();
for (MyNode s : first) {
buf.append(s).append(" ");
}
buf.append("==> ");
for (MyNode s : second) {
buf.append(s).append(" ");
}
buf.append("\n");
}
buf.append("requiredirect\n\n");
for (OrderedPair<Set<MyNode>> o : requiredRulesSpecs) {
Set<MyNode> first = o.getFirst();
Set<MyNode> second = o.getSecond();
for (MyNode s : first) {
buf.append(s).append(" ");
}
buf.append("==> ");
for (MyNode s : second) {
buf.append(s).append(" ");
}
buf.append("\n");
}
out.write(buf.toString());
out.flush();
}
private void resolveOneEdgeMax(Graph graph, Node x, Node y, boolean strong, Graph oldGraph) {
if (RandomUtil.getInstance().nextDouble() > 0.5) {
Node temp = x;
x = y;
y = temp;
}
TetradLogger.getInstance().log("info", "\nEDGE " + x + " --- " + y);
SortedMap<Double, String> scoreReports = new TreeMap<Double, String>();
List<Node> neighborsx = graph.getAdjacentNodes(x);
neighborsx.remove(y);
double max = Double.NEGATIVE_INFINITY;
boolean left = false;
boolean right = false;
DepthChoiceGenerator genx = new DepthChoiceGenerator(neighborsx.size(), neighborsx.size());
int[] choicex;
while ((choicex = genx.next()) != null) {
List<Node> condxMinus = GraphUtils.asList(choicex, neighborsx);
List<Node> condxPlus = new ArrayList<Node>(condxMinus);
condxPlus.add(y);
double xPlus = score(x, condxPlus);
double xMinus = score(x, condxMinus);
List<Node> neighborsy = graph.getAdjacentNodes(y);
neighborsy.remove(x);
DepthChoiceGenerator geny = new DepthChoiceGenerator(neighborsy.size(), neighborsy.size());
int[] choicey;
while ((choicey = geny.next()) != null) {
List<Node> condyMinus = GraphUtils.asList(choicey, neighborsy);
// List<Node> parentsY = oldGraph.getParents(y);
// parentsY.remove(x);
// if (!condyMinus.containsAll(parentsY)) {
// continue;
// }
List<Node> condyPlus = new ArrayList<Node>(condyMinus);
condyPlus.add(x);
double yPlus = score(y, condyPlus);
double yMinus = score(y, condyMinus);
// Checking them all at once is expensive but avoids lexical ordering problems in the
// algorithm.
if (normal(y, condyPlus)
|| normal(x, condxMinus)
|| normal(x, condxPlus)
|| normal(y, condyMinus)) {
continue;
}
double delta = 0.0;
if (strong) {
if (yPlus <= xPlus + delta && xMinus <= yMinus + delta) {
double score = combinedScore(xPlus, yMinus);
if (yPlus <= yMinus + delta && xMinus <= xPlus + delta) {
StringBuilder builder = new StringBuilder();
builder.append("\nStrong " + y + "->" + x + " " + score);
builder.append("\n Parents(" + x + ") = " + condxMinus);
builder.append("\n Parents(" + y + ") = " + condyMinus);
scoreReports.put(-score, builder.toString());
if (score > max) {
max = score;
left = true;
right = false;
}
} else {
StringBuilder builder = new StringBuilder();
builder.append("\nNo directed edge " + x + "--" + y + " " + score);
builder.append("\n Parents(" + x + ") = " + condxMinus);
builder.append("\n Parents(" + y + ") = " + condyMinus);
scoreReports.put(-score, builder.toString());
}
} else if (xPlus <= yPlus + delta && yMinus <= xMinus + delta) {
double score = combinedScore(yPlus, xMinus);
if (yMinus <= yPlus + delta && xPlus <= xMinus + delta) {
StringBuilder builder = new StringBuilder();
builder.append("\nStrong " + x + "->" + y + " " + score);
builder.append("\n Parents(" + x + ") = " + condxMinus);
builder.append("\n Parents(" + y + ") = " + condyMinus);
scoreReports.put(-score, builder.toString());
if (score > max) {
max = score;
left = false;
right = true;
}
} else {
StringBuilder builder = new StringBuilder();
builder.append("\nNo directed edge " + x + "--" + y + " " + score);
builder.append("\n Parents(" + x + ") = " + condxMinus);
builder.append("\n Parents(" + y + ") = " + condyMinus);
scoreReports.put(-score, builder.toString());
}
} else if (yPlus <= xPlus + delta && yMinus <= xMinus + delta) {
double score = combinedScore(yPlus, xMinus);
StringBuilder builder = new StringBuilder();
builder.append("\nNo directed edge " + x + "--" + y + " " + score);
builder.append("\n Parents(" + x + ") = " + condxMinus);
builder.append("\n Parents(" + y + ") = " + condyMinus);
scoreReports.put(-score, builder.toString());
} else if (xPlus <= yPlus + delta && xMinus <= yMinus + delta) {
double score = combinedScore(yPlus, xMinus);
StringBuilder builder = new StringBuilder();
builder.append("\nNo directed edge " + x + "--" + y + " " + score);
builder.append("\n Parents(" + x + ") = " + condxMinus);
builder.append("\n Parents(" + y + ") = " + condyMinus);
scoreReports.put(-score, builder.toString());
}
} else {
if (yPlus <= xPlus + delta && xMinus <= yMinus + delta) {
double score = combinedScore(xPlus, yMinus);
StringBuilder builder = new StringBuilder();
builder.append("\nWeak " + y + "->" + x + " " + score);
builder.append("\n Parents(" + x + ") = " + condxMinus);
builder.append("\n Parents(" + y + ") = " + condyMinus);
scoreReports.put(-score, builder.toString());
if (score > max) {
max = score;
left = true;
right = false;
}
} else if (xPlus <= yPlus + delta && yMinus <= xMinus + delta) {
double score = combinedScore(yPlus, xMinus);
StringBuilder builder = new StringBuilder();
builder.append("\nWeak " + x + "->" + y + " " + score);
builder.append("\n Parents(" + x + ") = " + condxMinus);
builder.append("\n Parents(" + y + ") = " + condyMinus);
scoreReports.put(-score, builder.toString());
if (score > max) {
max = score;
left = false;
right = true;
}
} else if (yPlus <= xPlus + delta && yMinus <= xMinus + delta) {
double score = combinedScore(yPlus, xMinus);
StringBuilder builder = new StringBuilder();
builder.append("\nNo directed edge " + x + "--" + y + " " + score);
builder.append("\n Parents(" + x + ") = " + condxMinus);
builder.append("\n Parents(" + y + ") = " + condyMinus);
scoreReports.put(-score, builder.toString());
} else if (xPlus <= yPlus + delta && xMinus <= yMinus + delta) {
double score = combinedScore(yPlus, xMinus);
StringBuilder builder = new StringBuilder();
builder.append("\nNo directed edge " + x + "--" + y + " " + score);
builder.append("\n Parents(" + x + ") = " + condxMinus);
builder.append("\n Parents(" + y + ") = " + condyMinus);
scoreReports.put(-score, builder.toString());
}
}
}
}
for (double score : scoreReports.keySet()) {
TetradLogger.getInstance().log("info", scoreReports.get(score));
}
graph.removeEdges(x, y);
if (left) {
graph.addDirectedEdge(y, x);
}
if (right) {
graph.addDirectedEdge(x, y);
}
if (!graph.isAdjacentTo(x, y)) {
graph.addUndirectedEdge(x, y);
}
}
public GeneralizedSemPm(SemPm semPm) {
this(semPm.getGraph());
// Write down equations.
try {
List<Node> variableNodes = getVariableNodes();
for (int i = 0; i < variableNodes.size(); i++) {
Node node = variableNodes.get(i);
List<Node> parents = getVariableParents(node);
StringBuilder buf = new StringBuilder();
for (int j = 0; j < parents.size(); j++) {
if (!(variableNodes.contains(parents.get(j)))) {
continue;
}
Node parent = parents.get(j);
Parameter _parameter = semPm.getParameter(parent, node);
String parameter = _parameter.getName();
Set<Node> nodes = new HashSet<>();
nodes.add(node);
referencedParameters.put(parameter, nodes);
buf.append(parameter);
buf.append("*");
buf.append(parents.get(j).getName());
setParameterExpression(parameter, "Split(-1.5, -.5, .5, 1.5)");
setStartsWithParametersTemplate(parameter.substring(0, 1), "Split(-1.5, -.5, .5, 1.5)");
setStartsWithParametersEstimationInitializaationTemplate(
parameter.substring(0, 1), "Split(-1.5, -.5, .5, 1.5)");
if (j < parents.size() - 1) {
buf.append(" + ");
}
}
if (buf.toString().trim().length() != 0) {
buf.append(" + ");
}
buf.append(errorNodes.get(i));
setNodeExpression(node, buf.toString());
}
for (Node node : variableNodes) {
Parameter _parameter = semPm.getParameter(node, node);
String parameter = _parameter.getName();
String distributionFormula = "N(0," + parameter + ")";
setNodeExpression(getErrorNode(node), distributionFormula);
setParameterExpression(parameter, "U(0, 1)");
setStartsWithParametersTemplate(parameter.substring(0, 1), "U(0, 1)");
setStartsWithParametersEstimationInitializaationTemplate(
parameter.substring(0, 1), "U(0, 1)");
}
variableNames = new ArrayList<>();
for (Node _node : variableNodes) variableNames.add(_node.getName());
for (Node _node : errorNodes) variableNames.add(_node.getName());
} catch (ParseException e) {
throw new IllegalStateException("Parse error in constructing initial model.", e);
}
}