commons-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From GitBox <...@apache.org>
Subject [GitHub] kinow commented on a change in pull request #103: TEXT-126: Adding Sorensen-Dice similarity algoritham
Date Sun, 03 Mar 2019 05:22:59 GMT
kinow commented on a change in pull request #103: TEXT-126: Adding Sorensen-Dice similarity
algoritham
URL: https://github.com/apache/commons-text/pull/103#discussion_r261852025
 
 

 ##########
 File path: src/main/java/org/apache/commons/text/similarity/SorensenDicesSimilarity.java
 ##########
 @@ -0,0 +1,118 @@
+/*
+ * 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.
+ */
+package org.apache.commons.text.similarity;
+
+import java.util.HashSet;
+import java.util.Set;
+
+/**
+ * A similarity algorithm indicating the percentage of matched characters
+ * between two character sequences.
+ *
+ * <p>
+ * The Sørensen–Dice coefficient is a statistic used for comparing the
+ * similarity of two samples. It was independently developed by the botanists
+ * Thorvald Sørensen and Lee Raymond Dice, who published in 1948 and 1945
+ * respectively. The index is known by several other names, especially
+ * Sørensen–Dice index, Sørensen index and Dice's coefficient. Other variations
+ * include the "similarity coefficient" or "index", such as Dice similarity
+ * coefficient (DSC).
+ * </p>
+ *
+ * <p>
+ * This implementation is based on the Sørensen–Dice similarity algorithm from
+ * <a href=
+ * "http://en.wikipedia.org/wiki/S%C3%B8rensen%E2%80%93Dice_coefficient">
+ * http://en.wikipedia.org/wiki/S%C3%B8rensen%E2%80%93Dice_coefficient</a>.
+ *
+ *
+ * </p>
+ *
+ * @since 1.7
+ */
+public class SorensenDicesSimilarity implements SimilarityScore<Double> {
+
+    /**
+     *
+     * <pre>
+     * similarity.apply("", "")                     = 1.0
+     * similarity.apply("foo", "foo")               = 1.0
+     * similarity.apply("foo", "foo ")              = 0.8
+     * similarity.apply("foo", "foo ")              = 0.66
+     * similarity.apply("foo", " foo ")             = 0.66
+     * similarity.apply("foo", " foo")              = 0.66
+     * similarity.apply("", "a")                    = 0.0
+     * similarity.apply("aaapppp", "")              = 0.0
+     * similarity.apply("frog", "fog")              = 0.4
+     * similarity.apply("fly", "ant")               = 0.0
+     * similarity.apply("elephant", "hippo")        = 0.0
+     * similarity.apply("hippo", "elephant")        = 0.0
+     * similarity.apply("hippo", "zzzzzzzz")        = 0.0
+     * similarity.apply("hello", "hallo")           = 0.5
+     * similarity.apply("ABC Corporation", "ABC Corp") = 0.7
+     * similarity.apply("D N H Enterprises Inc", "D &amp; H Enterprises, Inc.") = 0.74
+     * similarity.apply("My Gym Children's Fitness Center", "My Gym. Childrens Fitness")
= 0.81
+     * similarity.apply("PENNSYLVANIA", "PENNCISYLVNIA") = 0.69
+     * </pre>
+     *
+     * @param left  the first CharSequence, must not be null
+     * @param right the second CharSequence, must not be null
+     * @return result similarity
+     * @throws IllegalArgumentException if either CharSequence input is {@code null}
+     */
+    @Override
+    public Double apply(final CharSequence left, final CharSequence right) {
+
+        if (left == null || right == null) {
+            throw new IllegalArgumentException("CharSequences must not be null");
+        }
+
+        if (left.equals(right)) {
+            return 1d;
+        }
+
+        if (left.length() == 0 || right.length() == 0) {
+            return 0d;
+        }
+
+        Set<String> nLeft = createBigrams(left);
+        Set<String> nRight = createBigrams(right);
+
+        final int total = nLeft.size() + nRight.size();
+        nLeft.retainAll(nRight);
 
 Review comment:
   It may not be necessary to call `retainAll` I think? We want to `intersection` being the
number of elements common to both vectors/sets/sides/etc.
   
   Indeed, calling `retainAll` does the trick here. But I believe `createBigrams` is returning
`HashSet`, which is backed - as far as I can tell - by a `HashMap`, where get/put should be
quite good, but not so much delete operations (I remember looking at one implementation some
time ago, and it had so many checks in place :eyes: ).
   
   So _I think_ we can achieve the same by just filtering the collections and counting the
total number of elements common to both sets?
   
   e.g.
   
   ```diff
   diff --git a/src/main/java/org/apache/commons/text/similarity/SorensenDicesSimilarity.java
b/src/main/java/org/apache/commons/text/similarity/SorensenDicesSimilarity.java
   index d4df1d7..c79023a 100644
   --- a/src/main/java/org/apache/commons/text/similarity/SorensenDicesSimilarity.java
   +++ b/src/main/java/org/apache/commons/text/similarity/SorensenDicesSimilarity.java
   @@ -18,6 +18,7 @@ package org.apache.commons.text.similarity;
    
    import java.util.HashSet;
    import java.util.Set;
   +import java.util.stream.Collectors;
    
    /**
     * A similarity algorithm indicating the percentage of matched characters
   @@ -93,8 +94,9 @@ public class SorensenDicesSimilarity implements SimilarityScore<Double>
{
            Set<String> nRight = createBigrams(right);
    
            final int total = nLeft.size() + nRight.size();
   -        nLeft.retainAll(nRight);
   -        final int intersection = nLeft.size();
   +        final long intersection = nLeft.stream()
   +                .filter(nRight::contains)
   +                .collect(Collectors.counting());
    
            return (2.0d * intersection) / total;
        }
   ```
   
   I believe this way we take advantage of Java 8 syntax, and avoid creating a new data structure,
or performing multiple checks. The tests worked, but being quite late Sunday for me, almost
time to call it a day, and my last comment in this pull request, I may be saying something
totally silly. So no hard feelings if anyone corrects me :+1: 

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services

Mime
View raw message