Menu
▾
▴
[DL-Learner SVN] SF.net SVN: dl-learner:[3879] trunk
|
From: <jen...@us...> - 2012-11-29 15:21:53
|
Revision: 3879
http://dl-learner.svn.sourceforge.net/dl-learner/?rev=3879&view=rev
Author: jenslehmann
Date: 2012-11-29 15:21:42 +0000 (Thu, 29 Nov 2012)
Log Message:
-----------
documentation update
Modified Paths:
--------------
trunk/components-core/src/main/java/org/dllearner/learningproblems/ClassLearningProblem.java
trunk/components-core/src/main/java/org/dllearner/learningproblems/PosNegLPStandard.java
trunk/examples/father.conf
trunk/interfaces/doc/configOptions.html
Modified: trunk/components-core/src/main/java/org/dllearner/learningproblems/ClassLearningProblem.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/learningproblems/ClassLearningProblem.java 2012-11-29 14:57:34 UTC (rev 3878)
+++ trunk/components-core/src/main/java/org/dllearner/learningproblems/ClassLearningProblem.java 2012-11-29 15:21:42 UTC (rev 3879)
@@ -1,5 +1,5 @@
/**
- * Copyright (C) 2007-2011, Jens Lehmann
+ * Copyright (C) 2007-2012, Jens Lehmann
*
* This file is part of DL-Learner.
*
@@ -19,7 +19,6 @@
package org.dllearner.learningproblems;
-import java.util.Collection;
import java.util.Collections;
import java.util.Iterator;
import java.util.LinkedList;
@@ -34,12 +33,7 @@
import org.dllearner.core.ComponentAnn;
import org.dllearner.core.ComponentInitException;
import org.dllearner.core.ComponentManager;
-import org.dllearner.core.options.BooleanConfigOption;
-import org.dllearner.core.options.CommonConfigOptions;
-import org.dllearner.core.options.ConfigOption;
-import org.dllearner.core.options.DoubleConfigOption;
-import org.dllearner.core.options.StringConfigOption;
-import org.dllearner.core.options.URLConfigOption;
+import org.dllearner.core.config.ConfigOption;
import org.dllearner.core.owl.Axiom;
import org.dllearner.core.owl.Description;
import org.dllearner.core.owl.EquivalentClassesAxiom;
@@ -64,22 +58,26 @@
private long nanoStartTime;
private int maxExecutionTimeInSeconds = 10;
- // TODO: config option
+ @ConfigOption(name = "classToDescribe", description="class of which a description should be learned", required=true)
private NamedClass classToDescribe;
private List<Individual> classInstances;
private TreeSet<Individual> classInstancesSet;
private boolean equivalence = true;
-// private ClassLearningProblemConfigurator configurator;
- // approximation of accuracy
+
+ @ConfigOption(name = "approxDelta", description = "The Approximate Delta", defaultValue = "0.05", required = false)
private double approxDelta = 0.05;
+ @ConfigOption(name = "useApproximations", description = "Use Approximations", defaultValue = "false", required = false)
private boolean useApproximations;
// factor for higher weight on recall (needed for subclass learning)
private double coverageFactor;
+ @ConfigOption(name = "betaSC", description="beta index for F-measure in super class learning", required=false, defaultValue="3.0")
private double betaSC = 3.0;
+
+ @ConfigOption(name = "betaEq", description="beta index for F-measure in definition learning", required=false, defaultValue="1.0")
private double betaEq = 1.0;
// instances of super classes excluding instances of the class itself
@@ -89,8 +87,12 @@
// specific variables for generalised F-measure
private TreeSet<Individual> negatedClassInstances;
+ @ConfigOption(name = "accuracyMethod", description = "Specifies, which method/function to use for computing accuracy. Available measues are \"pred_acc\" (predictive accuracy), \"fmeasure\" (F measure), \"generalised_fmeasure\" (generalised F-Measure according to Fanizzi and d'Amato).",defaultValue = "pred_acc")
+ private String accuracyMethod = "pred_acc";
+
private HeuristicType heuristic = HeuristicType.AMEASURE;
+ @ConfigOption(name = "checkConsistency", description = "whether to check for consistency of suggestions (when added to ontology)", required=false, defaultValue="true")
private boolean checkConsistency = true;
public ClassLearningProblem() {
@@ -106,6 +108,7 @@
// configurator = new ClassLearningProblemConfigurator(this);
}
+ /*
public static Collection<ConfigOption<?>> createConfigOptions() {
Collection<ConfigOption<?>> options = new LinkedList<ConfigOption<?>>();
URLConfigOption classToDescribeOption = new URLConfigOption("classToDescribe", "class of which a description should be learned", null, true, false);
@@ -130,6 +133,7 @@
options.add(betaEq);
return options;
}
+ */
public static String getName() {
return "class learning problem";
@@ -140,18 +144,18 @@
// classToDescribe = new NamedClass(configurator.getClassToDescribe().toString());
// useApproximations = configurator.getUseApproximations();
-// String accM = configurator.getAccuracyMethod();
-// if(accM.equals("standard")) {
-// heuristic = HeuristicType.AMEASURE;
-// } else if(accM.equals("fmeasure")) {
-// heuristic = HeuristicType.FMEASURE;
-// } else if(accM.equals("generalised_fmeasure")) {
-// heuristic = HeuristicType.GEN_FMEASURE;
-// } else if(accM.equals("jaccard")) {
-// heuristic = HeuristicType.JACCARD;
-// } else if(accM.equals("pred_acc")) {
-// heuristic = HeuristicType.PRED_ACC;
-// }
+ String accM = accuracyMethod;
+ if(accM.equals("standard")) {
+ heuristic = HeuristicType.AMEASURE;
+ } else if(accM.equals("fmeasure")) {
+ heuristic = HeuristicType.FMEASURE;
+ } else if(accM.equals("generalised_fmeasure")) {
+ heuristic = HeuristicType.GEN_FMEASURE;
+ } else if(accM.equals("jaccard")) {
+ heuristic = HeuristicType.JACCARD;
+ } else if(accM.equals("pred_acc")) {
+ heuristic = HeuristicType.PRED_ACC;
+ }
if(useApproximations && heuristic.equals(HeuristicType.PRED_ACC)) {
System.err.println("Approximating predictive accuracy is an experimental feature. USE IT AT YOUR OWN RISK. If you consider to use it for anything serious, please extend the unit tests at org.dllearner.test.junit.HeuristicTests first to verify that it works.");
Modified: trunk/components-core/src/main/java/org/dllearner/learningproblems/PosNegLPStandard.java
===================================================================
--- trunk/components-core/src/main/java/org/dllearner/learningproblems/PosNegLPStandard.java 2012-11-29 14:57:34 UTC (rev 3878)
+++ trunk/components-core/src/main/java/org/dllearner/learningproblems/PosNegLPStandard.java 2012-11-29 15:21:42 UTC (rev 3879)
@@ -66,9 +66,9 @@
@ConfigOption(name = "useApproximations", description = "Use Approximations", defaultValue = "false", required = false)
private boolean useApproximations;
- @ConfigOption(name = "accuracyMethod", description = "Specifies, which method/function to use for computing accuracy.",defaultValue = "predacc", propertyEditorClass = StringTrimmerEditor.class)
- private String accuracyMethod = "predacc";
-
+ @ConfigOption(name = "accuracyMethod", description = "Specifies, which method/function to use for computing accuracy. Available measues are \"pred_acc\" (predictive accuracy), \"fmeasure\" (F measure), \"generalised_fmeasure\" (generalised F-Measure according to Fanizzi and d'Amato).",defaultValue = "predacc", propertyEditorClass = StringTrimmerEditor.class)
+ private String accuracyMethod = "pred_acc";
+
// private boolean useFMeasure;
private boolean useOldDIGOptions = false;
Modified: trunk/examples/father.conf
===================================================================
--- trunk/examples/father.conf 2012-11-29 14:57:34 UTC (rev 3878)
+++ trunk/examples/father.conf 2012-11-29 15:21:42 UTC (rev 3879)
@@ -24,5 +24,6 @@
lp.negativeExamples = { "ex:heinz", "ex:anna", "ex:michelle" }
// create learning algorithm to run
-alg1.type = "celoe"
-alg2.type = "pceloe"
+alg.type = "celoe"
+alg.maxExecutionTimeInSeconds = 1
+
Modified: trunk/interfaces/doc/configOptions.html
===================================================================
--- trunk/interfaces/doc/configOptions.html 2012-11-29 14:57:34 UTC (rev 3878)
+++ trunk/interfaces/doc/configOptions.html 2012-11-29 15:21:42 UTC (rev 3879)
@@ -65,27 +65,36 @@
<p>short name: bruteForce<br />version: 0.8<br />implements: LearningAlgorithm, ClassExpressionLearningAlgorithm<br /></p>This component does not have configuration options.</div>
<div class="LearningAlgorithm ClassExpressionLearningAlgorithm"><a name="org.dllearner.algorithms.celoe.CELOE"><h2>CELOE</h2></a>
<p>short name: celoe<br />version: 1.0<br />implements: LearningAlgorithm, ClassExpressionLearningAlgorithm<br />description: CELOE is an adapted and extended version of the OCEL algorithm applied for the ontology engineering use case. See http://jens-lehmann.org/files/2011/celoe.pdf for reference.<br /></p><table id="hor-minimalist-a"><thead><tr><th>option name</th><th>description</th><th>type</th><th>default value</th><th>required?</th></tr></thead><tbody>
+<tr><td>maxExecutionTimeInSecondsAfterImprovement</td><td>maximum execution of the algorithm in seconds</td><td> int</td><td>0</td><td> false</td></tr>
<tr><td>startClass</td><td>You can specify a start class for the algorithm. To do this, you have to use Manchester OWL syntax without using prefixes.</td><td> Description</td><td>owl:Thing</td><td> false</td></tr>
-<tr><td>maxClassExpressionTests</td><td>The maximum number of candidate hypothesis the algorithm is allowed to test (0 = no limit). The algorithm will stop afterwards. (The real number of tests can be slightly higher, because this criterion usually won't be checked after each single test.)</td><td> int</td><td>0</td><td> false</td></tr>
-<tr><td>maxClassExpressionTestsAfterImprovement</td><td>The maximum number of candidate hypothesis the algorithm is allowed after an improvement in accuracy (0 = no limit). The algorithm will stop afterwards. (The real number of tests can be slightly higher, because this criterion usually won't be checked after each single test.)</td><td> int</td><td>0</td><td> false</td></tr>
-<tr><td>maxExecutionTimeInSecondsAfterImprovement</td><td>maximum execution of the algorithm in seconds</td><td> int</td><td>0</td><td> false</td></tr>
-<tr><td>replaceSearchTree</td><td>specifies whether to replace the search tree in the log file after each run or append the new search tree</td><td> boolean</td><td>false</td><td> false</td></tr>
-<tr><td>useMinimizer</td><td>Specifies whether returned expressions should be minimised by removing those parts, which are not needed. (Basically the minimiser tries to find the shortest expression which is equivalent to the learned expression). Turning this feature off may improve performance.</td><td> boolean</td><td>true</td><td> false</td></tr>
+<tr><td>noisePercentage</td><td>the (approximated) percentage of noise within the examples</td><td> double</td><td>0.0</td><td> false</td></tr>
+<tr><td>terminateOnNoiseReached</td><td>specifies whether to terminate when noise criterion is met</td><td> boolean</td><td>false</td><td> false</td></tr>
+<tr><td>writeSearchTree</td><td>specifies whether to write a search tree</td><td> boolean</td><td>false</td><td> false</td></tr>
<tr><td>filterDescriptionsFollowingFromKB</td><td>If true, then the results will not contain suggestions, which already follow logically from the knowledge base. Be careful, since this requires a potentially expensive consistency check for candidate solutions.</td><td> boolean</td><td>false</td><td> false</td></tr>
-<tr><td>stopOnFirstDefinition</td><td>algorithm will terminate immediately when a correct definition is found</td><td> boolean</td><td>false</td><td> false</td></tr>
-<tr><td>reuseExistingDescription</td><td>If true, the algorithm tries to find a good starting point close to an existing definition/super class of the given class in the knowledge base.</td><td> boolean</td><td>false</td><td> false</td></tr>
-<tr><td>terminateOnNoiseReached</td><td>specifies whether to terminate when noise criterion is met</td><td> boolean</td><td>false</td><td> false</td></tr>
-<tr><td>noisePercentage</td><td>the (approximated) percentage of noise within the examples</td><td> double</td><td>0.0</td><td> false</td></tr>
<tr><td>maxDepth</td><td>maximum depth of description</td><td> double</td><td>7</td><td> false</td></tr>
<tr><td>searchTreeFile</td><td>file to use for the search tree</td><td> String</td><td>log/searchTree.txt</td><td> false</td></tr>
<tr><td>maxExecutionTimeInSeconds</td><td>maximum execution of the algorithm in seconds</td><td> int</td><td>10</td><td> false</td></tr>
-<tr><td>writeSearchTree</td><td>specifies whether to write a search tree</td><td> boolean</td><td>false</td><td> false</td></tr>
+<tr><td>useMinimizer</td><td>Specifies whether returned expressions should be minimised by removing those parts, which are not needed. (Basically the minimiser tries to find the shortest expression which is equivalent to the learned expression). Turning this feature off may improve performance.</td><td> boolean</td><td>true</td><td> false</td></tr>
+<tr><td>maxClassExpressionTestsAfterImprovement</td><td>The maximum number of candidate hypothesis the algorithm is allowed after an improvement in accuracy (0 = no limit). The algorithm will stop afterwards. (The real number of tests can be slightly higher, because this criterion usually won't be checked after each single test.)</td><td> int</td><td>0</td><td> false</td></tr>
+<tr><td>reuseExistingDescription</td><td>If true, the algorithm tries to find a good starting point close to an existing definition/super class of the given class in the knowledge base.</td><td> boolean</td><td>false</td><td> false</td></tr>
+<tr><td>replaceSearchTree</td><td>specifies whether to replace the search tree in the log file after each run or append the new search tree</td><td> boolean</td><td>false</td><td> false</td></tr>
+<tr><td>stopOnFirstDefinition</td><td>algorithm will terminate immediately when a correct definition is found</td><td> boolean</td><td>false</td><td> false</td></tr>
+<tr><td>maxClassExpressionTests</td><td>The maximum number of candidate hypothesis the algorithm is allowed to test (0 = no limit). The algorithm will stop afterwards. (The real number of tests can be slightly higher, because this criterion usually won't be checked after each single test.)</td><td> int</td><td>0</td><td> false</td></tr>
+<tr><td>singleSuggestionMode</td><td>Use this if you are interested in only one suggestion and your learning problem has many (more than 1000) examples.</td><td> boolean</td><td>false</td><td> false</td></tr>
<tr><td>maxNrOfResults</td><td>Sets the maximum number of results one is interested in. (Setting this to a lower value may increase performance as the learning algorithm has to store/evaluate/beautify less descriptions).</td><td> int</td><td>10</td><td> false</td></tr>
-<tr><td>singleSuggestionMode</td><td>Use this if you are interested in only one suggestion and your learning problem has many (more than 1000) examples.</td><td> boolean</td><td>false</td><td> false</td></tr>
</tbody></table>
</div>
<div class="LearningProblem"><a name="org.dllearner.learningproblems.ClassLearningProblem"><h2>ClassLearningProblem</h2></a>
-<p>short name: clp<br />version: 0.6<br />implements: LearningProblem<br /></p>This component does not have configuration options.</div>
+<p>short name: clp<br />version: 0.6<br />implements: LearningProblem<br /></p><table id="hor-minimalist-a"><thead><tr><th>option name</th><th>description</th><th>type</th><th>default value</th><th>required?</th></tr></thead><tbody>
+<tr><td>accuracyMethod</td><td>Specifies, which method/function to use for computing accuracy. Available measues are "pred_acc" (predictive accuracy), "fmeasure" (F measure), "generalised_fmeasure" (generalised F-Measure according to Fanizzi and d'Amato).</td><td> String</td><td>pred_acc</td><td> false</td></tr>
+<tr><td>classToDescribe</td><td>class of which a description should be learned</td><td> NamedClass</td><td></td><td> true</td></tr>
+<tr><td>checkConsistency</td><td>whether to check for consistency of suggestions (when added to ontology)</td><td> boolean</td><td>true</td><td> false</td></tr>
+<tr><td>approxDelta</td><td>The Approximate Delta</td><td> double</td><td>0.05</td><td> false</td></tr>
+<tr><td>betaSC</td><td>beta index for F-measure in super class learning</td><td> double</td><td>3.0</td><td> false</td></tr>
+<tr><td>betaEq</td><td>beta index for F-measure in definition learning</td><td> double</td><td>1.0</td><td> false</td></tr>
+<tr><td>useApproximations</td><td>Use Approximations</td><td> boolean</td><td>false</td><td> false</td></tr>
+</tbody></table>
+</div>
<div class="LearningAlgorithm ClassExpressionLearningAlgorithm"><a name="org.dllearner.algorithms.el.ELLearningAlgorithmDisjunctive"><h2>Disjunctive ELTL</h2></a>
<p>short name: deltl<br />version: 0.5<br />implements: LearningAlgorithm, ClassExpressionLearningAlgorithm<br />description: Disjunctive ELTL is an algorithm based on the refinement operator in http://jens-lehmann.org/files/2009/el_ilp.pdf with support for disjunctions.<br /></p><table id="hor-minimalist-a"><thead><tr><th>option name</th><th>description</th><th>type</th><th>default value</th><th>required?</th></tr></thead><tbody>
<tr><td>tryFullCoverage</td><td>If yes, then the algorithm tries to cover all positive examples. Note that while this improves accuracy on the testing set, it may lead to overfitting.</td><td> boolean</td><td>false</td><td> false</td></tr>
@@ -105,8 +114,8 @@
<p>short name: fuzzyPosNeg<br />version: 0.2<br />implements: LearningProblem<br /></p>This component does not have configuration options.</div>
<div class="KnowledgeSource"><a name="org.dllearner.kb.KBFile"><h2>KB File</h2></a>
<p>short name: kbfile<br />version: 0.8<br />implements: KnowledgeSource<br /></p><table id="hor-minimalist-a"><thead><tr><th>option name</th><th>description</th><th>type</th><th>default value</th><th>required?</th></tr></thead><tbody>
+<tr><td>url</td><td>URL pointer to the KB file</td><td> String</td><td></td><td> false</td></tr>
<tr><td>fileName</td><td>relative or absolute path to KB file</td><td> String</td><td></td><td> false</td></tr>
-<tr><td>url</td><td>URL pointer to the KB file</td><td> String</td><td></td><td> false</td></tr>
</tbody></table>
</div>
<div class="OtherComponent"><a name="org.dllearner.algorithms.celoe.OEHeuristicRuntime"><h2>OEHeuristicRuntime</h2></a>
@@ -116,8 +125,8 @@
</div>
<div class="ReasonerComponent"><a name="org.dllearner.reasoning.OWLAPIReasoner"><h2>OWL API Reasoner</h2></a>
<p>short name: oar<br />version: 0.8<br />implements: ReasonerComponent<br /></p><table id="hor-minimalist-a"><thead><tr><th>option name</th><th>description</th><th>type</th><th>default value</th><th>required?</th></tr></thead><tbody>
+<tr><td>reasonerType</td><td>The name of the OWL APIReasoner to use {"fact", "hermit", "owllink", "pellet", "elk", "cel"}</td><td> String</td><td>pellet</td><td> false</td></tr>
<tr><td>owlLinkURL</td><td>The URL to the owl server</td><td> String</td><td></td><td> false</td></tr>
-<tr><td>reasonerType</td><td>The name of the OWL APIReasoner to use {"fact", "hermit", "owllink", "pellet", "elk", "cel"}</td><td> String</td><td>pellet</td><td> false</td></tr>
</tbody></table>
</div>
<div class="LearningAlgorithm ClassExpressionLearningAlgorithm"><a name="org.dllearner.algorithms.ocel.OCEL"><h2>OWL Class Expression Learner</h2></a>
@@ -126,45 +135,45 @@
<p>short name: owlfile<br />version: 0.9<br />implements: KnowledgeSource<br /></p>This component does not have configuration options.</div>
<div class="LearningAlgorithm ClassExpressionLearningAlgorithm"><a name="org.dllearner.algorithms.celoe.PCELOE"><h2>PCELOE</h2></a>
<p>short name: pceloe<br />version: 1.0<br />implements: LearningAlgorithm, ClassExpressionLearningAlgorithm<br />description: CELOE is an adapted and extended version of the OCEL algorithm applied for the ontology engineering use case. See http://jens-lehmann.org/files/2011/celoe.pdf for reference.<br /></p><table id="hor-minimalist-a"><thead><tr><th>option name</th><th>description</th><th>type</th><th>default value</th><th>required?</th></tr></thead><tbody>
+<tr><td>maxExecutionTimeInSecondsAfterImprovement</td><td>maximum execution of the algorithm in seconds</td><td> int</td><td>0</td><td> false</td></tr>
<tr><td>startClass</td><td>You can specify a start class for the algorithm. To do this, you have to use Manchester OWL syntax without using prefixes.</td><td> Description</td><td>owl:Thing</td><td> false</td></tr>
-<tr><td>maxClassExpressionTests</td><td>The maximum number of candidate hypothesis the algorithm is allowed to test (0 = no limit). The algorithm will stop afterwards. (The real number of tests can be slightly higher, because this criterion usually won't be checked after each single test.)</td><td> int</td><td>0</td><td> false</td></tr>
-<tr><td>maxClassExpressionTestsAfterImprovement</td><td>The maximum number of candidate hypothesis the algorithm is allowed after an improvement in accuracy (0 = no limit). The algorithm will stop afterwards. (The real number of tests can be slightly higher, because this criterion usually won't be checked after each single test.)</td><td> int</td><td>0</td><td> false</td></tr>
-<tr><td>maxExecutionTimeInSecondsAfterImprovement</td><td>maximum execution of the algorithm in seconds</td><td> int</td><td>0</td><td> false</td></tr>
-<tr><td>replaceSearchTree</td><td>specifies whether to replace the search tree in the log file after each run or append the new search tree</td><td> boolean</td><td>false</td><td> false</td></tr>
-<tr><td>useMinimizer</td><td>Specifies whether returned expressions should be minimised by removing those parts, which are not needed. (Basically the minimiser tries to find the shortest expression which is equivalent to the learned expression). Turning this feature off may improve performance.</td><td> boolean</td><td>true</td><td> false</td></tr>
-<tr><td>filterDescriptionsFollowingFromKB</td><td>If true, then the results will not contain suggestions, which already follow logically from the knowledge base. Be careful, since this requires a potentially expensive consistency check for candidate solutions.</td><td> boolean</td><td>false</td><td> false</td></tr>
+<tr><td>noisePercentage</td><td>the (approximated) percentage of noise within the examples</td><td> double</td><td>0.0</td><td> false</td></tr>
<tr><td>nrOfThreads</td><td>number of threads running in parallel</td><td> int</td><td>2</td><td> false</td></tr>
-<tr><td>reuseExistingDescription</td><td>If true, the algorithm tries to find a good starting point close to an existing definition/super class of the given class in the knowledge base.</td><td> boolean</td><td>false</td><td> false</td></tr>
<tr><td>terminateOnNoiseReached</td><td>specifies whether to terminate when noise criterion is met</td><td> boolean</td><td>false</td><td> false</td></tr>
-<tr><td>noisePercentage</td><td>the (approximated) percentage of noise within the examples</td><td> double</td><td>0.0</td><td> false</td></tr>
+<tr><td>writeSearchTree</td><td>specifies whether to write a search tree</td><td> boolean</td><td>false</td><td> false</td></tr>
+<tr><td>filterDescriptionsFollowingFromKB</td><td>If true, then the results will not contain suggestions, which already follow logically from the knowledge base. Be careful, since this requires a potentially expensive consistency check for candidate solutions.</td><td> boolean</td><td>false</td><td> false</td></tr>
<tr><td>maxDepth</td><td>maximum depth of description</td><td> double</td><td>7</td><td> false</td></tr>
<tr><td>searchTreeFile</td><td>file to use for the search tree</td><td> String</td><td>log/searchTree.txt</td><td> false</td></tr>
<tr><td>maxExecutionTimeInSeconds</td><td>maximum execution of the algorithm in seconds</td><td> int</td><td>10</td><td> false</td></tr>
-<tr><td>writeSearchTree</td><td>specifies whether to write a search tree</td><td> boolean</td><td>false</td><td> false</td></tr>
+<tr><td>useMinimizer</td><td>Specifies whether returned expressions should be minimised by removing those parts, which are not needed. (Basically the minimiser tries to find the shortest expression which is equivalent to the learned expression). Turning this feature off may improve performance.</td><td> boolean</td><td>true</td><td> false</td></tr>
+<tr><td>maxClassExpressionTestsAfterImprovement</td><td>The maximum number of candidate hypothesis the algorithm is allowed after an improvement in accuracy (0 = no limit). The algorithm will stop afterwards. (The real number of tests can be slightly higher, because this criterion usually won't be checked after each single test.)</td><td> int</td><td>0</td><td> false</td></tr>
+<tr><td>reuseExistingDescription</td><td>If true, the algorithm tries to find a good starting point close to an existing definition/super class of the given class in the knowledge base.</td><td> boolean</td><td>false</td><td> false</td></tr>
+<tr><td>replaceSearchTree</td><td>specifies whether to replace the search tree in the log file after each run or append the new search tree</td><td> boolean</td><td>false</td><td> false</td></tr>
+<tr><td>maxClassExpressionTests</td><td>The maximum number of candidate hypothesis the algorithm is allowed to test (0 = no limit). The algorithm will stop afterwards. (The real number of tests can be slightly higher, because this criterion usually won't be checked after each single test.)</td><td> int</td><td>0</td><td> false</td></tr>
+<tr><td>singleSuggestionMode</td><td>Use this if you are interested in only one suggestion and your learning problem has many (more than 1000) examples.</td><td> boolean</td><td>false</td><td> false</td></tr>
<tr><td>maxNrOfResults</td><td>Sets the maximum number of results one is interested in. (Setting this to a lower value may increase performance as the learning algorithm has to store/evaluate/beautify less descriptions).</td><td> int</td><td>10</td><td> false</td></tr>
-<tr><td>singleSuggestionMode</td><td>Use this if you are interested in only one suggestion and your learning problem has many (more than 1000) examples.</td><td> boolean</td><td>false</td><td> false</td></tr>
</tbody></table>
</div>
<div class="LearningProblem"><a name="org.dllearner.learningproblems.PosNegLPStandard"><h2>PosNegLPStandard</h2></a>
<p>short name: posNegStandard<br />version: 0.8<br />implements: LearningProblem<br /></p><table id="hor-minimalist-a"><thead><tr><th>option name</th><th>description</th><th>type</th><th>default value</th><th>required?</th></tr></thead><tbody>
<tr><td>approxDelta</td><td>The Approximate Delta</td><td> double</td><td>0.05</td><td> false</td></tr>
-<tr><td>accuracyMethod</td><td>Specifies, which method/function to use for computing accuracy.</td><td> String</td><td>predacc</td><td> false</td></tr>
+<tr><td>accuracyMethod</td><td>Specifies, which method/function to use for computing accuracy. Available measues are "pred_acc" (predictive accuracy), "fmeasure" (F measure), "generalised_fmeasure" (generalised F-Measure according to Fanizzi and d'Amato).</td><td> String</td><td>predacc</td><td> false</td></tr>
<tr><td>useApproximations</td><td>Use Approximations</td><td> boolean</td><td>false</td><td> false</td></tr>
</tbody></table>
</div>
<div class="LearningProblem"><a name="org.dllearner.learningproblems.PosNegLPStrict"><h2>PosNegLPStrict</h2></a>
<p>short name: posNegStrict<br />version: 0.8<br />implements: LearningProblem<br /></p><table id="hor-minimalist-a"><thead><tr><th>option name</th><th>description</th><th>type</th><th>default value</th><th>required?</th></tr></thead><tbody>
+<tr><td>errorPenalty</td><td>penalty for errors (example can be inferred to belong to the negated concept class)</td><td> double</td><td>3</td><td> false</td></tr>
<tr><td>accuracyPenalty</td><td>penalty for incorrectness (example belongs neither to concept nor its negation)</td><td> double</td><td>1</td><td> false</td></tr>
-<tr><td>errorPenalty</td><td>penalty for errors (example can be inferred to belong to the negated concept class)</td><td> double</td><td>3</td><td> false</td></tr>
</tbody></table>
</div>
<div class="LearningAlgorithm ClassExpressionLearningAlgorithm"><a name="org.dllearner.algorithms.RandomGuesser"><h2>Random Guesser</h2></a>
<p>short name: randomGuesser<br />version: 0.8<br />implements: LearningAlgorithm, ClassExpressionLearningAlgorithm<br /></p>This component does not have configuration options.</div>
<div class="KnowledgeSource"><a name="org.dllearner.kb.SparqlEndpointKS"><h2>SPARQL endpoint</h2></a>
<p>short name: sparql<br />version: 0.2<br />implements: KnowledgeSource<br /></p><table id="hor-minimalist-a"><thead><tr><th>option name</th><th>description</th><th>type</th><th>default value</th><th>required?</th></tr></thead><tbody>
+<tr><td>url</td><td>no description available</td><td> URL</td><td></td><td> true</td></tr>
<tr><td>namedGraphs</td><td>no description available</td><td> List</td><td>[]</td><td> false</td></tr>
<tr><td>defaultGraphs</td><td>no description available</td><td> List</td><td>[]</td><td> false</td></tr>
-<tr><td>url</td><td>no description available</td><td> URL</td><td></td><td> true</td></tr>
</tbody></table>
</div>
<div class="KnowledgeSource"><a name="org.dllearner.kb.sparql.SparqlKnowledgeSource"><h2>SPARQL endpoint fragment</h2></a>
@@ -201,14 +210,14 @@
</div>
<div class="KnowledgeSource"><a name="org.dllearner.kb.sparql.simple.SparqlSimpleExtractor"><h2>efficient SPARQL fragment extractor</h2></a>
<p>short name: sparqls<br />version: 0.1<br />implements: KnowledgeSource<br /></p><table id="hor-minimalist-a"><thead><tr><th>option name</th><th>description</th><th>type</th><th>default value</th><th>required?</th></tr></thead><tbody>
+<tr><td>aboxfilter</td><td>Filter for the tbox, can use variable ?s, ?p amd ?o</td><td> String</td><td></td><td> false</td></tr>
+<tr><td>defaultGraphURI</td><td>default graph URI</td><td> String</td><td></td><td> true</td></tr>
+<tr><td>ontologySchemaUrls</td><td>List of Ontology Schema URLs</td><td> List</td><td></td><td> true</td></tr>
+<tr><td>sparqlQuery</td><td>Sparql Query</td><td> String</td><td></td><td> false</td></tr>
<tr><td>tboxfilter</td><td>Filter for the tbox, can use variable ?example and ?class</td><td> String</td><td></td><td> false</td></tr>
+<tr><td>recursionDepth</td><td>recursion depth</td><td> int</td><td></td><td> true</td></tr>
<tr><td>endpointURL</td><td>URL of the SPARQL endpoint</td><td> String</td><td></td><td> true</td></tr>
-<tr><td>ontologySchemaUrls</td><td>List of Ontology Schema URLs</td><td> List</td><td></td><td> true</td></tr>
-<tr><td>aboxfilter</td><td>Filter for the tbox, can use variable ?s, ?p amd ?o</td><td> String</td><td></td><td> false</td></tr>
<tr><td>instances</td><td>List of the instances to use</td><td> List</td><td></td><td> true</td></tr>
-<tr><td>sparqlQuery</td><td>Sparql Query</td><td> String</td><td></td><td> false</td></tr>
-<tr><td>defaultGraphURI</td><td>default graph URI</td><td> String</td><td></td><td> true</td></tr>
-<tr><td>recursionDepth</td><td>recursion depth</td><td> int</td><td></td><td> true</td></tr>
</tbody></table>
</div>
<div class="LearningAlgorithm AxiomLearningAlgorithm"><a name="org.dllearner.algorithms.properties.EquivalentDataPropertyAxiomLearner"><h2>equivalent dataproperty axiom learner</h2></a>
@@ -245,10 +254,10 @@
<div class="OtherComponent"><a name="org.dllearner.algorithms.ocel.MultiHeuristic"><h2>multiple criteria heuristic</h2></a>
<p>short name: multiheuristic<br />version: 0.7<br />implements: OtherComponent<br /></p><table id="hor-minimalist-a"><thead><tr><th>option name</th><th>description</th><th>type</th><th>default value</th><th>required?</th></tr></thead><tbody>
<tr><td>expansionPenaltyFactor</td><td>no description available</td><td> double</td><td>0.02</td><td> false</td></tr>
-<tr><td>gainBonusFactor</td><td>no description available</td><td> double</td><td>0.5</td><td> false</td></tr>
+<tr><td>negativeWeight</td><td>no description available</td><td> double</td><td>1.0</td><td> false</td></tr>
<tr><td>startNodeBonus</td><td>no description available</td><td> double</td><td>0.1</td><td> false</td></tr>
<tr><td>negationPenalty</td><td>no description available</td><td> int</td><td>0</td><td> false</td></tr>
-<tr><td>negativeWeight</td><td>no description available</td><td> double</td><td>1.0</td><td> false</td></tr>
+<tr><td>gainBonusFactor</td><td>no description available</td><td> double</td><td>0.5</td><td> false</td></tr>
<tr><td>nodeChildPenalty</td><td>no description available</td><td> double</td><td>0.0001</td><td> false</td></tr>
</tbody></table>
</div>
@@ -271,19 +280,19 @@
<p>short name: posonlylp<br />version: 0.6<br />implements: LearningProblem<br /></p>This component does not have configuration options.</div>
<div class="RefinementOperator"><a name="org.dllearner.refinementoperators.RhoDRDown"><h2>rho refinement operator</h2></a>
<p>short name: rho<br />version: 0.8<br />implements: RefinementOperator<br /></p><table id="hor-minimalist-a"><thead><tr><th>option name</th><th>description</th><th>type</th><th>default value</th><th>required?</th></tr></thead><tbody>
+<tr><td>instanceBasedDisjoints</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
+<tr><td>useStringDatatypes</td><td>no description available</td><td> boolean</td><td>false</td><td> false</td></tr>
+<tr><td>useNegation</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
<tr><td>useAllConstructor</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
+<tr><td>disjointChecks</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
+<tr><td>applyExistsFilter</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
<tr><td>useHasValueConstructor</td><td>no description available</td><td> boolean</td><td>false</td><td> false</td></tr>
-<tr><td>instanceBasedDisjoints</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
+<tr><td>useBooleanDatatypes</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
+<tr><td>useExistsConstructor</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
<tr><td>dropDisjuncts</td><td>no description available</td><td> boolean</td><td>false</td><td> false</td></tr>
<tr><td>applyAllFilter</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
<tr><td>useDoubleDatatypes</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
-<tr><td>useNegation</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
-<tr><td>useStringDatatypes</td><td>no description available</td><td> boolean</td><td>false</td><td> false</td></tr>
-<tr><td>useBooleanDatatypes</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
-<tr><td>useExistsConstructor</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
<tr><td>useCardinalityRestrictions</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
-<tr><td>applyExistsFilter</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
-<tr><td>disjointChecks</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
</tbody></table>
</div>
<div class="LearningAlgorithm AxiomLearningAlgorithm ClassExpressionLearningAlgorithm"><a name="org.dllearner.algorithms.SimpleSubclassLearner"><h2>simple subclass learner</h2></a>
@@ -298,19 +307,19 @@
</div>
<div class="RefinementOperator"><a name="org.dllearner.refinementoperators.SynchronizedRhoDRDown"><h2>synchronized rho refinement operator</h2></a>
<p>short name: syncrho<br />version: 0.8<br />implements: RefinementOperator<br /></p><table id="hor-minimalist-a"><thead><tr><th>option name</th><th>description</th><th>type</th><th>default value</th><th>required?</th></tr></thead><tbody>
+<tr><td>instanceBasedDisjoints</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
+<tr><td>useStringDatatypes</td><td>no description available</td><td> boolean</td><td>false</td><td> false</td></tr>
+<tr><td>useNegation</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
<tr><td>useAllConstructor</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
+<tr><td>disjointChecks</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
+<tr><td>applyExistsFilter</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
<tr><td>useHasValueConstructor</td><td>no description available</td><td> boolean</td><td>false</td><td> false</td></tr>
-<tr><td>instanceBasedDisjoints</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
+<tr><td>useBooleanDatatypes</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
+<tr><td>useExistsConstructor</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
<tr><td>dropDisjuncts</td><td>no description available</td><td> boolean</td><td>false</td><td> false</td></tr>
<tr><td>applyAllFilter</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
<tr><td>useDoubleDatatypes</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
-<tr><td>useNegation</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
-<tr><td>useStringDatatypes</td><td>no description available</td><td> boolean</td><td>false</td><td> false</td></tr>
-<tr><td>useBooleanDatatypes</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
-<tr><td>useExistsConstructor</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
<tr><td>useCardinalityRestrictions</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
-<tr><td>applyExistsFilter</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
-<tr><td>disjointChecks</td><td>no description available</td><td> boolean</td><td>true</td><td> false</td></tr>
</tbody></table>
</div>
<div class="LearningAlgorithm AxiomLearningAlgorithm"><a name="org.dllearner.algorithms.properties.TransitiveObjectPropertyAxiomLearner"><h2>transitive objectproperty axiom learner</h2></a>
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
