001/** 002 * Copyright (C) 2007 - 2016, Jens Lehmann 003 * 004 * This file is part of DL-Learner. 005 * 006 * DL-Learner is free software; you can redistribute it and/or modify 007 * it under the terms of the GNU General Public License as published by 008 * the Free Software Foundation; either version 3 of the License, or 009 * (at your option) any later version. 010 * 011 * DL-Learner is distributed in the hope that it will be useful, 012 * but WITHOUT ANY WARRANTY; without even the implied warranty of 013 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 014 * GNU General Public License for more details. 015 * 016 * You should have received a copy of the GNU General Public License 017 * along with this program. If not, see <http://www.gnu.org/licenses/>. 018 */ 019package org.dllearner.learningproblems; 020 021import com.google.common.collect.Sets; 022import com.google.common.collect.Sets.SetView; 023import org.apache.log4j.Logger; 024import org.dllearner.accuracymethods.AccMethodApproximate; 025import org.dllearner.accuracymethods.AccMethodPredAcc; 026import org.dllearner.accuracymethods.AccMethodTwoValued; 027import org.dllearner.core.AbstractClassExpressionLearningProblem; 028import org.dllearner.core.AbstractReasonerComponent; 029import org.dllearner.core.ComponentInitException; 030import org.dllearner.core.config.ConfigOption; 031import org.dllearner.reasoning.SPARQLReasoner; 032import org.dllearner.utilities.Helper; 033import org.semanticweb.owlapi.model.OWLIndividual; 034import org.semanticweb.owlapi.model.OWLNamedIndividual; 035import org.springframework.beans.factory.annotation.Autowired; 036 037import java.util.Set; 038import java.util.TreeSet; 039 040/** 041 * @author Jens Lehmann 042 * 043 */ 044public abstract class PosNegLP extends AbstractClassExpressionLearningProblem<ScorePosNeg<OWLNamedIndividual>> { 045 protected static final Logger logger = Logger.getLogger(PosNegLP.class); 046 047 @ConfigOption(description = "list of positive examples", required = true) 048 protected Set<OWLIndividual> positiveExamples = new TreeSet<>(); 049 @ConfigOption(description = "list of negative examples", required = true) 050 protected Set<OWLIndividual> negativeExamples = new TreeSet<>(); 051 protected Set<OWLIndividual> allExamples = new TreeSet<>(); 052 053 @ConfigOption(description = "\"Specifies whether to use retrieval or instance checks for testing a concept. - NO LONGER FULLY SUPPORTED.",defaultValue = "false") 054 private boolean useRetrievalForClassification = false; 055 @ConfigOption(description = "Percent Per Length Unit", defaultValue = "0.05", required = false) 056 private double percentPerLengthUnit = 0.05; 057 058 @ConfigOption(description = "Specifies, which method/function to use for computing accuracy. Available measues are \"PRED_ACC\" (predictive accuracy), \"FMEASURE\" (F measure), \"GEN_FMEASURE\" (generalised F-Measure according to Fanizzi and d'Amato).", 059 defaultValue = "PRED_ACC") 060 protected AccMethodTwoValued accuracyMethod; 061 062 public PosNegLP(){ 063 064 } 065 066 public PosNegLP(AbstractReasonerComponent reasoningService) { 067 super(reasoningService); 068 } 069 070 /* 071 * (non-Javadoc) 072 * 073 * @see org.dllearner.core.Component#init() 074 */ 075 @Override 076 public void init() throws ComponentInitException { 077 ExampleLoader exampleLoaderHelper = this.getExampleLoaderHelper(); 078 if (exampleLoaderHelper != null && !exampleLoaderHelper.isInitialized()) { 079 logger.info("Loading examples by expression"); 080 exampleLoaderHelper.setPosNegLP(this); 081 exampleLoaderHelper.init(); 082 } 083 // check if some positive examples have been set 084 if(positiveExamples.isEmpty()) { 085 throw new ComponentInitException("No positive examples have been set."); 086 } 087 088 // check if some negative examples have been set and give warning if not 089 if(negativeExamples.isEmpty()) { 090 logger.warn("No negative examples have been set, but you decided to use the positive-negative learning" 091 + "problem. We recommend to use the positive-only learning problem for the case of no negative examples instead."); 092 } 093 094 // check if there is some overlap between positive and negative examples and give warning 095 // in that case 096 SetView<OWLIndividual> overlap = Sets.intersection(positiveExamples, negativeExamples); 097 if(!overlap.isEmpty()) { 098 logger.warn("You declared some individuals as both positive and negative examples."); 099 } 100 101 allExamples = Sets.union(positiveExamples, negativeExamples); 102 103 if (accuracyMethod == null) { 104 accuracyMethod = new AccMethodPredAcc(true); 105 } 106 if (accuracyMethod instanceof AccMethodApproximate) { 107 ((AccMethodApproximate)accuracyMethod).setReasoner(reasoner); 108 } 109 110 // sanity check whether examples are contained in KB 111 Helper.checkIndividuals(reasoner, allExamples); 112 113 initialized = true; 114 } 115 116 public Set<OWLIndividual> getNegativeExamples() { 117 return negativeExamples; 118 } 119 120 public Set<OWLIndividual> getPositiveExamples() { 121 return positiveExamples; 122 } 123 124 public void setNegativeExamples(Set<OWLIndividual> set) { 125 this.negativeExamples=set; 126 } 127 128 public void setPositiveExamples(Set<OWLIndividual> set) { 129 this.positiveExamples=set; 130 } 131 132 public double getPercentPerLengthUnit() { 133 return percentPerLengthUnit; 134 } 135 136 public void setPercentPerLengthUnit(double percentPerLengthUnit) { 137 this.percentPerLengthUnit = percentPerLengthUnit; 138 } 139 140 public boolean isUseRetrievalForClassification() { 141 return useRetrievalForClassification; 142 } 143 144 public void setUseRetrievalForClassification(boolean useRetrievalForClassification) { 145 this.useRetrievalForClassification = useRetrievalForClassification; 146 } 147 148 public AccMethodTwoValued getAccuracyMethod() { 149 return accuracyMethod; 150 } 151 152 @Autowired(required=false) 153 public void setAccuracyMethod(AccMethodTwoValued accuracyMethod) { 154 this.accuracyMethod = accuracyMethod; 155 } 156 157}