Adding new kernel metrics to the ctakes-ytex concept similarity service.

ctakes-ytex/src/main/java/org/apache/ctakes/ytex/kernel/metric/BraunBlanquetMetric.java

@@ -0,0 +1,98 @@
+package org.apache.ctakes.ytex.kernel.metric;
+
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements.  See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership.  The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License.  You may obtain a copy of the License at
+ *
+ *   http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing,
+ * software distributed under the License is distributed on an
+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+ * KIND, either express or implied.  See the License for the
+ * specific language governing permissions and limitations
+ * under the License.
+ */
+
+import java.util.Map;
+
+import org.apache.commons.logging.Log;
+import org.apache.commons.logging.LogFactory;
+
+/**
+ * Compute the Dice metric 
+ * 
+ * 
+ * Described in:
+ * https://www.sciencedirect.com/science/article/pii/S1532046411000645 Tbl 3, eqn 11
+ * 
+ * @author painter
+ * 
+ */
+public class BraunBlanquetMetric extends BaseSimilarityMetric {
+
+	private static final Log log = LogFactory.getLog(LinMetric.class);
+	private boolean intrinsicIC = true;
+	private boolean validCG = false;
+	private String rootConcept = simSvc.getConceptGraph().getRoot();
+
+	public boolean isIntrinsicIC() {
+		return intrinsicIC;
+	}
+
+	public void setIntrinsicIC(boolean intrinsicIC) {
+		this.intrinsicIC = intrinsicIC;
+	}
+
+	@Override
+	public double similarity(String concept1, String concept2,
+			Map<String, Double> conceptFilter, SimilarityInfo simInfo) {
+		// don't bother if the concept graph is null
+		if (!validCG)
+			return 0d;
+
+		// Compute the IC values for each concept
+		double ic1 = simSvc.getIC(concept1, this.intrinsicIC);
+		double ic2 = simSvc.getIC(concept2, this.intrinsicIC);
+
+		// Get the LCS with the lowest IC score
+		double lcsIC = initLcsIC(concept1, concept2, conceptFilter, simInfo,
+				this.intrinsicIC);
+
+		// if the corpus IC is 0 and the concept is not the root, then we don't
+		// have any IC on the concept and can't measure similarity - return 0
+		if (!intrinsicIC && ic1 == 0 && !rootConcept.equals(concept1))
+			return 0d;
+
+		if (!intrinsicIC && ic2 == 0 && !rootConcept.equals(concept2))
+			return 0d;
+
+		// Compute the Braun-Blanquet score
+		// we just need one of these to be greater than zero
+		if ( ic1 > 0 || ic2 > 0 ) {
+
+			// max of the individual IC's
+			double denom = ic1;
+			if ( ic2 > ic1 ) denom = ic2;
+			double sim = (lcsIC) / ( denom );
+			return sim;	
+		} else {
+			return 0d;
+		}
+	}
+
+	public BraunBlanquetMetric(ConceptSimilarityService simSvc, boolean intrinsicIC) {
+		super(simSvc);
+		this.intrinsicIC = intrinsicIC;
+		this.validCG = simSvc.getConceptGraph() != null;
+		if (!this.intrinsicIC && validCG) {
+			rootConcept = simSvc.getConceptGraph().getRoot();
+		}
+	}
+
+}

ctakes-ytex/src/main/java/org/apache/ctakes/ytex/kernel/metric/ConceptSimilarityService.java

@@ -29,10 +29,14 @@
 public interface ConceptSimilarityService {
 
 	public enum SimilarityMetricEnum {
-		LCH(false, false), INTRINSIC_LCH(true, false), LIN(false, true), INTRINSIC_LIN(
-				true, false), PATH(false, false), INTRINSIC_PATH(true, false), JACCARD(
-				true, false), SOKAL(true, false), RADA(false, false), INTRINSIC_RADA(
-				true, false), WUPALMER(false, false), PAGERANK(false, false);
+		        LCH(false, false), INTRINSIC_LCH(true, false), 
+		        LIN(false, true), INTRINSIC_LIN(true, false), PATH(false, false), 
+		        INTRINSIC_PATH(true, false), JACCARD(true, false), SOKAL(true, false), 
+		        RADA(false, false), INTRINSIC_RADA(true, false), WUPALMER(false, false), 
+		        PAGERANK(false, false), RESNIK(false, false), INTRINSIC_RESNIK(true, false),
+				FAITH(false, false), INTRINSIC_FAITH(true, false), DICE(false, true),
+				SIMPSON(false, true), BRAUN_BLANQUET(false, true), OCHIAI(false, true);
+
 		boolean intrinsicIC = false;
 		boolean corpusIC = false;
 
@@ -201,4 +205,4 @@ public List<ConceptPairSimilarity> similarity(
 			Map<String, Double> conceptFilter, boolean lcs);
 
 	public abstract int getDepth(String concept);
-}
+}

ctakes-ytex/src/main/java/org/apache/ctakes/ytex/kernel/metric/ConceptSimilarityServiceImpl.java

@@ -856,6 +856,24 @@ private void initSimilarityMetricMap() {
 					new JaccardMetric(this));
 			this.similarityMetricMap.put(SimilarityMetricEnum.WUPALMER,
 					new WuPalmerMetric(this));
+			this.similarityMetricMap.put(SimilarityMetricEnum.INTRINSIC_RESNIK,
+					new ResnikMetric(this, true));
+			this.similarityMetricMap.put(SimilarityMetricEnum.RESNIK,
+					new ResnikMetric(this, false));
+			this.similarityMetricMap.put(SimilarityMetricEnum.INTRINSIC_FAITH,
+					new FaithMetric(this, true));
+			this.similarityMetricMap.put(SimilarityMetricEnum.FAITH,
+					new FaithMetric(this, false));
+			this.similarityMetricMap.put(SimilarityMetricEnum.DICE,
+					new DiceMetric(this, true));
+			this.similarityMetricMap.put(SimilarityMetricEnum.SIMPSON,
+					new SimpsonMetric(this, true));
+			this.similarityMetricMap.put(SimilarityMetricEnum.BRAUN_BLANQUET,
+					new BraunBlanquetMetric(this, true));
+			this.similarityMetricMap.put(SimilarityMetricEnum.OCHIAI,
+					new OchiaiMetric(this, true));
+
+
 		} else {
 			this.similarityMetricMap.put(SimilarityMetricEnum.PAGERANK,
 					new PageRankMetric(this, this.getPageRankService()));

ctakes-ytex/src/main/java/org/apache/ctakes/ytex/kernel/metric/DiceMetric.java

@@ -0,0 +1,93 @@
+package org.apache.ctakes.ytex.kernel.metric;
+
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements.  See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership.  The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License.  You may obtain a copy of the License at
+ *
+ *   http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing,
+ * software distributed under the License is distributed on an
+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+ * KIND, either express or implied.  See the License for the
+ * specific language governing permissions and limitations
+ * under the License.
+ */
+
+import java.util.Map;
+
+import org.apache.commons.logging.Log;
+import org.apache.commons.logging.LogFactory;
+
+/**
+ * Compute the Dice metric 
+ * 
+ * 
+ * Described in:
+ * https://www.sciencedirect.com/science/article/pii/S1532046411000645 Tbl 3, eqn 8
+ * 
+ * @author painter
+ * 
+ */
+public class DiceMetric extends BaseSimilarityMetric {
+
+	private static final Log log = LogFactory.getLog(LinMetric.class);
+	private boolean intrinsicIC = true;
+	private boolean validCG = false;
+	private String rootConcept = simSvc.getConceptGraph().getRoot();
+
+	public boolean isIntrinsicIC() {
+		return intrinsicIC;
+	}
+
+	public void setIntrinsicIC(boolean intrinsicIC) {
+		this.intrinsicIC = intrinsicIC;
+	}
+
+	@Override
+	public double similarity(String concept1, String concept2,
+			Map<String, Double> conceptFilter, SimilarityInfo simInfo) {
+		// don't bother if the concept graph is null
+		if (!validCG)
+			return 0d;
+
+		// Compute the IC values for each concept
+		double ic1 = simSvc.getIC(concept1, this.intrinsicIC);
+		double ic2 = simSvc.getIC(concept2, this.intrinsicIC);
+
+		// Get the LCS with the lowest IC score
+		double lcsIC = initLcsIC(concept1, concept2, conceptFilter, simInfo,
+				this.intrinsicIC);
+
+		// if the corpus IC is 0 and the concept is not the root, then we don't
+		// have any IC on the concept and can't measure similarity - return 0
+		if (!intrinsicIC && ic1 == 0 && !rootConcept.equals(concept1))
+			return 0d;
+
+		if (!intrinsicIC && ic2 == 0 && !rootConcept.equals(concept2))
+			return 0d;
+
+		// Compute the Dice score
+		if ( ic1 > 0 || ic2 > 0 ) {
+			double sim = (2.0 * lcsIC) / ( ic1 + ic2 );
+			return sim;	
+		} else {
+			return 0d;
+		}
+	}
+
+	public DiceMetric(ConceptSimilarityService simSvc, boolean intrinsicIC) {
+		super(simSvc);
+		this.intrinsicIC = intrinsicIC;
+		this.validCG = simSvc.getConceptGraph() != null;
+		if (!this.intrinsicIC && validCG) {
+			rootConcept = simSvc.getConceptGraph().getRoot();
+		}
+	}
+
+}

ctakes-ytex/src/main/java/org/apache/ctakes/ytex/kernel/metric/FaithMetric.java

@@ -0,0 +1,96 @@
+package org.apache.ctakes.ytex.kernel.metric;
+
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements.  See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership.  The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License.  You may obtain a copy of the License at
+ *
+ *   http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing,
+ * software distributed under the License is distributed on an
+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+ * KIND, either express or implied.  See the License for the
+ * specific language governing permissions and limitations
+ * under the License.
+ */
+
+import java.util.Map;
+
+import org.apache.commons.logging.Log;
+import org.apache.commons.logging.LogFactory;
+
+/**
+ * compute Faith score to provide functionality as found in UMLS::Similarity
+ * 
+ * UMLS::Similarity::faith.pm
+ * Module implementing the semantic relatedness measure described 
+ * by Pirro and Euzenat (2010)
+ * 
+ * Described in:
+ * https://inria.hal.science/hal-00793283/file/pirro2010b.pdf
+ * 
+ * @author painter
+ * 
+ */
+public class FaithMetric extends BaseSimilarityMetric {
+
+	private static final Log log = LogFactory.getLog(LinMetric.class);
+	private boolean intrinsicIC = true;
+	private boolean validCG = false;
+	private String rootConcept = simSvc.getConceptGraph().getRoot();
+
+	public boolean isIntrinsicIC() {
+		return intrinsicIC;
+	}
+
+	public void setIntrinsicIC(boolean intrinsicIC) {
+		this.intrinsicIC = intrinsicIC;
+	}
+
+	@Override
+	public double similarity(String concept1, String concept2,
+			Map<String, Double> conceptFilter, SimilarityInfo simInfo) {
+		// don't bother if the concept graph is null
+		if (!validCG)
+			return 0d;
+
+		// Compute the IC values for each concept
+		double ic1 = simSvc.getIC(concept1, this.intrinsicIC);
+		double ic2 = simSvc.getIC(concept2, this.intrinsicIC);
+
+		// Get the LCS with the lowest IC score
+		double lcsIC = initLcsIC(concept1, concept2, conceptFilter, simInfo,
+				this.intrinsicIC);
+
+		// if the corpus IC is 0 and the concept is not the root, then we don't
+		// have any IC on the concept and can't measure similarity - return 0
+		if (!intrinsicIC && ic1 == 0 && !rootConcept.equals(concept1))
+			return 0d;
+
+		if (!intrinsicIC && ic2 == 0 && !rootConcept.equals(concept2))
+			return 0d;
+
+		// Compute the faith score
+		if ( ic1 > 0 && ic2 > 0 ) {
+			double sim = (lcsIC) / ( ic1 + ic2 - lcsIC );
+			return sim;	
+		} else {
+			return 0d;
+		}
+	}
+
+	public FaithMetric(ConceptSimilarityService simSvc, boolean intrinsicIC) {
+		super(simSvc);
+		this.intrinsicIC = intrinsicIC;
+		this.validCG = simSvc.getConceptGraph() != null;
+		if (!this.intrinsicIC && validCG) {
+			rootConcept = simSvc.getConceptGraph().getRoot();
+		}
+	}
+
+}