use of dr.evomodel.operators.ImportancePruneAndRegraft in project beast-mcmc by beast-dev.
the class ImportancePruneAndRegraftTestProblem method testDoOperation.
/**
* Test method for {@link SimpleMCMCOperator#doOperation()}.
* @throws ImportException
* @throws IOException
*/
public void testDoOperation() throws IOException, ImportException {
// probability of picking (A,B) node is 1/(2n-3) = 1/7
// probability of swapping with D is 1/2
// total = 1/14
//probability of picking {D} node is 1/(2n-3) = 1/7
//probability of picking {A,B} is 1/5
// total = 1/35
//total = 1/14 + 1/35 = 7/70 = 0.1
System.out.println("Test 1: Forward");
String treeMatch = "(((D,C),(A,B)),E);";
int count = 0;
int reps = 1000000;
for (int i = 0; i < reps; i++) {
TreeModel treeModel = new TreeModel("treeModel", tree5);
ImportancePruneAndRegraft operator = new ImportancePruneAndRegraft(treeModel, 1.0, 0);
operator.doOperation();
String tree = TreeUtils.newickNoLengths(treeModel);
if (tree.equals(treeMatch)) {
count += 1;
}
}
double p_1 = (double) count / (double) reps;
System.out.println("Number of proposals:\t" + count);
System.out.println("Number of tries:\t" + reps);
System.out.println("Number of ratio:\t" + p_1);
System.out.println("Number of expected ratio:\t" + 0.1);
assertExpectation(0.1, p_1, reps);
// lets see what the backward probability is for the hastings ratio
// (((D:2.0,C:2.0):1.0,(A:1.0,B:1.0):2.0):1.0,E:4.0) -> ((((A,B),C),D),E)
// probability of picking (A,B) node is 1/(2n-3) = 1/7
// probability of swapping with D is 1/3
// total = 1/21
//probability of picking {D} node is 1/(2n-2) = 1/7
//probability of picking {A,B} is 1/4
// total = 1/28
//total = 1/21 + 1/28 = 7/84 = 0.08333333
System.out.println("Test 2: Backward");
treeMatch = "((((A,B),C),D),E);";
NewickImporter importer = new NewickImporter("(((D:2.0,C:2.0):1.0,(A:1.0,B:1.0):2.0):1.0,E:4.0);");
FlexibleTree tree5_2 = (FlexibleTree) importer.importTree(null);
count = 0;
for (int i = 0; i < reps; i++) {
TreeModel treeModel = new TreeModel("treeModel", tree5_2);
ImportancePruneAndRegraft operator = new ImportancePruneAndRegraft(treeModel, 1.0, 1);
operator.doOperation();
String tree = TreeUtils.newickNoLengths(treeModel);
if (tree.equals(treeMatch)) {
count += 1;
}
}
double p_2 = (double) count / (double) reps;
System.out.println("Number of proposals:\t" + count);
System.out.println("Number of tries:\t" + reps);
System.out.println("Number of ratio:\t" + p_2);
System.out.println("Number of expected ratio:\t" + 0.0833333);
assertExpectation(0.0833333, p_2, reps);
}
use of dr.evomodel.operators.ImportancePruneAndRegraft in project beast-mcmc by beast-dev.
the class ImportancePruneAndRegraftParser method parseXMLObject.
public Object parseXMLObject(XMLObject xo) throws XMLParseException {
TreeModel treeModel = (TreeModel) xo.getChild(TreeModel.class);
double weight = xo.getDoubleAttribute(MCMCOperator.WEIGHT);
int samples = xo.getIntegerAttribute("samples");
return new ImportancePruneAndRegraft(treeModel, weight, samples);
}
use of dr.evomodel.operators.ImportancePruneAndRegraft in project beast-mcmc by beast-dev.
the class ImportancePruneAndRegraftTestProblem method getOperatorSchedule.
public OperatorSchedule getOperatorSchedule(TreeModel treeModel) {
Parameter rootParameter = treeModel.createNodeHeightsParameter(true, false, false);
Parameter internalHeights = treeModel.createNodeHeightsParameter(false, true, false);
ImportancePruneAndRegraft operator = new ImportancePruneAndRegraft(treeModel, 1.0, 1);
ScaleOperator scaleOperator = new ScaleOperator(rootParameter, 0.75, CoercionMode.COERCION_ON, 1.0);
UniformOperator uniformOperator = new UniformOperator(internalHeights, 1.0);
OperatorSchedule schedule = new SimpleOperatorSchedule();
schedule.addOperator(operator);
schedule.addOperator(scaleOperator);
schedule.addOperator(uniformOperator);
return schedule;
}
Aggregations