Menu
▾
▴
[DL-Learner SVN] SF.net SVN: dl-learner:[1109] trunk/src/dl-learner/org/dllearner
|
From: <ku...@us...> - 2008-08-20 14:37:21
|
Revision: 1109
http://dl-learner.svn.sourceforge.net/dl-learner/?rev=1109&view=rev
Author: kurzum
Date: 2008-08-20 14:37:13 +0000 (Wed, 20 Aug 2008)
Log Message:
-----------
Modified Paths:
--------------
trunk/src/dl-learner/org/dllearner/kb/aquisitors/SparqlTupelAquisitorImproved.java
trunk/src/dl-learner/org/dllearner/scripts/WikipediaCategoryCleaner.java
trunk/src/dl-learner/org/dllearner/server/DLLearnerWS.java
trunk/src/dl-learner/org/dllearner/utilities/examples/AutomaticNegativeExampleFinderSPARQL.java
Added Paths:
-----------
trunk/src/dl-learner/org/dllearner/test/TestGetExampleBug.java
Modified: trunk/src/dl-learner/org/dllearner/kb/aquisitors/SparqlTupelAquisitorImproved.java
===================================================================
--- trunk/src/dl-learner/org/dllearner/kb/aquisitors/SparqlTupelAquisitorImproved.java 2008-08-20 14:22:56 UTC (rev 1108)
+++ trunk/src/dl-learner/org/dllearner/kb/aquisitors/SparqlTupelAquisitorImproved.java 2008-08-20 14:37:13 UTC (rev 1109)
@@ -122,9 +122,14 @@
@Override
public SortedSet<RDFNodeTuple> retrieveTuplesForClassesOnly(String uri){
+ int tmp = recursionDepth;
+ recursionDepth=4;
+ SortedSet<RDFNodeTuple> tmpSet = retrieveTupel(uri);
+ recursionDepth = tmp;
+ return tmpSet;
//getQuery
- String sparqlQueryString = sparqlQueryMaker.makeSubjectQueryUsingFilters(uri);
- return sparqlTasks.queryAsRDFNodeTuple(sparqlQueryString, PREDICATE, OBJECT);
+ //String sparqlQueryString = sparqlQueryMaker.makeSubjectQueryUsingFilters(uri);
+ //return sparqlTasks.queryAsRDFNodeTuple(sparqlQueryString, PREDICATE, OBJECT);
}
Modified: trunk/src/dl-learner/org/dllearner/scripts/WikipediaCategoryCleaner.java
===================================================================
--- trunk/src/dl-learner/org/dllearner/scripts/WikipediaCategoryCleaner.java 2008-08-20 14:22:56 UTC (rev 1108)
+++ trunk/src/dl-learner/org/dllearner/scripts/WikipediaCategoryCleaner.java 2008-08-20 14:37:13 UTC (rev 1109)
@@ -223,8 +223,8 @@
lsc.sparqlEndpoint = sparqlTasks.getSparqlEndpoint();
lsc.recursiondepth = 1;
- lsc.noisePercentage = 15;
- lsc.guaranteeXgoodDescriptions = 200;
+ lsc.noisePercentage = 20;
+ lsc.guaranteeXgoodDescriptions = 100;
lsc.maxExecutionTimeInSeconds = 50;
lsc.logLevel = "TRACE";
// lsc.searchTreeFile = "log/WikipediaCleaner.txt";
Modified: trunk/src/dl-learner/org/dllearner/server/DLLearnerWS.java
===================================================================
--- trunk/src/dl-learner/org/dllearner/server/DLLearnerWS.java 2008-08-20 14:22:56 UTC (rev 1108)
+++ trunk/src/dl-learner/org/dllearner/server/DLLearnerWS.java 2008-08-20 14:37:13 UTC (rev 1109)
@@ -681,15 +681,28 @@
@WebMethod
public String[] getNegativeExamples(int sessionID, int componentID,String[] positives, int results, String namespace) throws ClientNotKnownException
{
- SortedSet<String> set = new TreeSet<String>(Arrays.asList(positives));
+ int sparqlResultSetLimit = 500;
+ SortedSet<String> positiveSet = new TreeSet<String>(Arrays.asList(positives));
ClientState state = getState(sessionID);
Component component = state.getComponent(componentID);
SparqlKnowledgeSource ks=(SparqlKnowledgeSource)component;
SPARQLTasks task=ks.getSPARQLTasks();
- AutomaticNegativeExampleFinderSPARQL finder=new AutomaticNegativeExampleFinderSPARQL(set,task);
- //finder.makeNegativeExamplesFromRelatedInstances(set, namespace);
- finder.makeNegativeExamplesFromParallelClasses(set, 25);
+ AutomaticNegativeExampleFinderSPARQL finder=new AutomaticNegativeExampleFinderSPARQL(positiveSet,task);
+
+ finder.makeNegativeExamplesFromParallelClasses(positiveSet, sparqlResultSetLimit);
SortedSet<String> negExamples=finder.getNegativeExamples(results);
+ if(negExamples.isEmpty()){
+ finder.makeNegativeExamplesFromRelatedInstances(positiveSet, namespace);
+ negExamples = finder.getNegativeExamples(results);
+ if(negExamples.isEmpty()){
+ finder.makeNegativeExamplesFromSuperClassesOfInstances(positiveSet, sparqlResultSetLimit);
+ negExamples = finder.getNegativeExamples(results);
+ if(negExamples.isEmpty()) {
+ finder.makeNegativeExamplesFromRandomInstances();
+ negExamples = finder.getNegativeExamples(results);
+ }
+ }
+ }
return negExamples.toArray(new String[negExamples.size()]);
}
Added: trunk/src/dl-learner/org/dllearner/test/TestGetExampleBug.java
===================================================================
--- trunk/src/dl-learner/org/dllearner/test/TestGetExampleBug.java (rev 0)
+++ trunk/src/dl-learner/org/dllearner/test/TestGetExampleBug.java 2008-08-20 14:37:13 UTC (rev 1109)
@@ -0,0 +1,86 @@
+package org.dllearner.test;
+
+import java.net.URL;
+import java.util.SortedSet;
+import java.util.TreeSet;
+
+import org.apache.log4j.ConsoleAppender;
+import org.apache.log4j.Level;
+import org.apache.log4j.Logger;
+import org.apache.log4j.SimpleLayout;
+import org.dllearner.kb.sparql.SPARQLTasks;
+import org.dllearner.kb.sparql.SparqlEndpoint;
+import org.dllearner.utilities.examples.AutomaticNegativeExampleFinderSPARQL;
+
+public class TestGetExampleBug {
+
+ /**
+ * @param args
+ */
+ public static void main(String[] args) {
+ Logger logger = Logger.getRootLogger();
+ SimpleLayout layout = new SimpleLayout();
+ ConsoleAppender consoleAppender = new ConsoleAppender(layout);
+ logger.removeAllAppenders();
+ logger.addAppender(consoleAppender);
+ logger.setLevel(Level.TRACE);
+ //Logger.getLogger(SparqlQuery.class).setLevel(Level.DEBUG);
+
+ try {
+ String OntowikiUrl="http://localhost/ontowiki-0.8.5/service/sparql";
+
+ String posExamples = "http://3ba.se/conferences/FMILeipzig";
+ SortedSet<String> positiveSet = new TreeSet<String>();
+ positiveSet.add(posExamples);
+ SPARQLTasks st = new SPARQLTasks(new SparqlEndpoint(new URL(OntowikiUrl)));
+ AutomaticNegativeExampleFinderSPARQL ane =
+ new AutomaticNegativeExampleFinderSPARQL(positiveSet,st);
+ SortedSet<String> negExamples = new TreeSet<String>();
+ int results=100;
+ if(negExamples.isEmpty()){
+ //ane.makeNegativeExamplesFromRelatedInstances(positiveSet, "http://3ba.se/conferences/");
+ negExamples = ane.getNegativeExamples(results);
+ if(negExamples.isEmpty()){
+ ane.makeNegativeExamplesFromSuperClassesOfInstances(positiveSet, 500);
+ negExamples = ane.getNegativeExamples(results);
+ if(negExamples.isEmpty()) {
+ ane.makeNegativeExamplesFromRandomInstances();
+ negExamples = ane.getNegativeExamples(results);
+ }
+ }
+ }
+
+ System.out.println(ane.getNegativeExamples(100));
+ }catch (Exception e) {
+ e.printStackTrace();
+ }
+
+ /*$negExamples=$client->getNegativeExamples($id,$ksID,$posExamples,count($posExamples),"http://localhost/ontowiki/service/sparql");
+ $negExamples=$negExamples->item;
+ $client->setLearningProblem($id, "posNegDefinition");
+ $client->setPositiveExamples($id, $posExamples);
+ $client->setNegativeExamples($id, $negExamples);
+
+// choose refinement operator approach
+ $client->setLearningAlgorithm($id, "refexamples");
+// you can add the following to apply a config option to a component, e.g. ignore a concept
+// $client->applyConfigEntryStringArray($id, $la_id, "ignoredConcepts", array('http://example.com/father#male'));
+
+ $client->initAll($id);
+
+// learn concept
+ echo 'start learning ... ';
+// get only concept
+// $concept = $client->learn($id, "manchester");
+// get concept and additional information in JSON syntax
+ $concept = $client->learnDescriptionsEvaluated($id, 5);
+ echo 'OK <br />';
+ echo 'solution: <pre>' . $concept . '</pre>';
+
+ ?>*/
+
+
+
+ }
+
+}
Modified: trunk/src/dl-learner/org/dllearner/utilities/examples/AutomaticNegativeExampleFinderSPARQL.java
===================================================================
--- trunk/src/dl-learner/org/dllearner/utilities/examples/AutomaticNegativeExampleFinderSPARQL.java 2008-08-20 14:22:56 UTC (rev 1108)
+++ trunk/src/dl-learner/org/dllearner/utilities/examples/AutomaticNegativeExampleFinderSPARQL.java 2008-08-20 14:37:13 UTC (rev 1109)
@@ -19,6 +19,7 @@
*/
package org.dllearner.utilities.examples;
+import java.util.Iterator;
import java.util.SortedSet;
import java.util.TreeSet;
@@ -26,6 +27,7 @@
import org.dllearner.core.ComponentManager;
import org.dllearner.kb.sparql.SPARQLTasks;
import org.dllearner.utilities.datastructures.SetManipulation;
+import org.dllearner.utilities.owl.OWLVocabulary;
public class AutomaticNegativeExampleFinderSPARQL {
@@ -40,6 +42,7 @@
private SortedSet<String> fromRelated = new TreeSet<String>();
private SortedSet<String> fromSuperclasses = new TreeSet<String>();;
private SortedSet<String> fromParallelClasses = new TreeSet<String>();;
+ private SortedSet<String> fromRandom = new TreeSet<String>();;
private SortedSet<String> fromDomain = new TreeSet<String>();;
private SortedSet<String> fromRange = new TreeSet<String>();;
@@ -68,16 +71,20 @@
/**
* aggregates all collected neg examples
+ * CAVE: it is necessary to call one of the make functions before calling this
+ *
* @param neglimit size of negative Example set
* @param stable decides whether neg Examples are randomly picked, default false, faster for developing, since the cache can be used
* @return
*/
public SortedSet<String> getNegativeExamples(int neglimit, boolean stable ) {
-
SortedSet<String> negatives = new TreeSet<String>();
negatives.addAll(fromParallelClasses);
negatives.addAll(fromRelated);
negatives.addAll(fromSuperclasses);
+ if(negatives.isEmpty()) {
+ negatives.addAll(fromRandom);
+ }
logger.debug("neg Example size before shrinking: " + negatives.size());
if (stable) {
negatives = SetManipulation.stableShrink(negatives,neglimit);
@@ -90,6 +97,18 @@
}
+ public void makeNegativeExamplesFromRandomInstances() {
+ logger.debug("making random examples ");
+ String variable = "subject";
+ String sparqlQueryString="SELECT ?"+variable+" WHERE {" +
+ "?"+variable+" <" +OWLVocabulary.RDF_TYPE+">" + " ?o" +
+ "}";
+
+ fromRandom = sparqltasks.queryAsSet(sparqlQueryString, variable);
+ fromRandom.removeAll(fullPositiveSet);
+ logger.debug("|-negExample size from related: " + fromRelated.size());
+ }
+
/**
* makes neg ex from related instances, that take part in a role R(pos,neg)
* filters all objects, that don't use the given namespace
@@ -173,23 +192,48 @@
}
+
+
/**
+ * it gets the first class of an arbitrary instance and queries the superclasses of it,
+ * could be more elaborate.
+ * It is better to use makeNegativeExamplesFromSuperClasses
+ * @param positiveSet
+ * @param sparqlResultSetLimit
+ */
+ public void makeNegativeExamplesFromSuperClassesOfInstances(SortedSet<String> positiveSet,
+ int sparqlResultSetLimit) {
+ SortedSet<String> classes = new TreeSet<String>();
+ Iterator<String> instanceIter = positiveSet.iterator();
+ while(classes.isEmpty() && instanceIter.hasNext()) {
+ classes.addAll(sparqltasks.getClassesForInstance(instanceIter.next(), sparqlResultSetLimit));
+
+ }
+ makeNegativeExamplesFromSuperClasses(classes.first(), sparqlResultSetLimit);
+ }
+
+
+ public void makeNegativeExamplesFromSuperClasses(String concept, int sparqlResultSetLimit) {
+ makeNegativeExamplesFromSuperClasses( concept, sparqlResultSetLimit, 2);
+ }
+
+ /**
* if pos ex derive from one class, then neg ex are taken from a superclass
* @param concept
- * @param resultLimit
+ * @param sparqlResultSetLimit
*/
- public void makeNegativeExamplesFromSuperClasses(String concept, int resultLimit) {
+ public void makeNegativeExamplesFromSuperClasses(String concept, int sparqlResultSetLimit, int depth) {
concept = concept.replaceAll("\"", "");
// superClasses.add(concept.replace("\"", ""));
// logger.debug("before"+superClasses);
- SortedSet<String> superClasses = sparqltasks.getSuperClasses(concept, 4);
+ SortedSet<String> superClasses = sparqltasks.getSuperClasses(concept, depth);
logger.debug("making neg Examples from " + superClasses.size() + " superclasses");
for (String oneSuperClass : superClasses) {
logger.debug(oneSuperClass);
fromSuperclasses.addAll(sparqltasks.retrieveInstancesForClassDescription("\""
- + oneSuperClass + "\"", resultLimit));
+ + oneSuperClass + "\"", sparqlResultSetLimit));
}
this.fromSuperclasses.removeAll(fullPositiveSet);
@@ -197,17 +241,17 @@
}
@SuppressWarnings("unused")
- private void makeNegativeExamplesFromDomain(String role, int resultLimit){
+ private void makeNegativeExamplesFromDomain(String role, int sparqlResultSetLimit){
logger.debug("making Negative Examples from Domain of : "+role);
- fromDomain.addAll(sparqltasks.getDomainInstances(role, resultLimit));
+ fromDomain.addAll(sparqltasks.getDomainInstances(role, sparqlResultSetLimit));
fromDomain.removeAll(fullPositiveSet);
logger.debug("|-neg Example size from Domain: "+this.fromDomain.size());
}
@SuppressWarnings("unused")
- private void makeNegativeExamplesFromRange(String role, int resultLimit){
+ private void makeNegativeExamplesFromRange(String role, int sparqlResultSetLimit){
logger.debug("making Negative Examples from Range of : "+role);
- fromRange.addAll(sparqltasks.getRangeInstances(role, resultLimit));
+ fromRange.addAll(sparqltasks.getRangeInstances(role, sparqlResultSetLimit));
fromRange.removeAll(fullPositiveSet);
logger.debug("|-neg Example size from Range: "+fromRange.size());
}
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
