Altered knn and ee to use predefined neighbours and full test neighbours

time-series-machine-learning · Jun 20, 2019 · 07408d1 · 07408d1
1 parent 2fad5fd
commit 07408d1
Show file tree

Hide file tree

Showing 14 changed files with 381 additions and 358 deletions.
diff --git a/src/main/java/classifiers/distance_based/elastic_ensemble/Distribute.java b/src/main/java/classifiers/distance_based/elastic_ensemble/Distribute.java
@@ -8,6 +8,7 @@
 import weka.core.Instance;
 import weka.core.Instances;
 
+import javax.rmi.CORBA.Util;
 import java.io.*;
 import java.nio.channels.FileChannel;
 import java.nio.channels.FileLock;
@@ -38,28 +39,30 @@ public static void main(String[] args) throws
     Instances[] data = sampleDataset(datasetDirPath, datasetName, seed);
     Instances train = data[0];
     Instances test = data[1];
-    ElasticEnsemble ee = new ElasticEnsemble();
-    ee.setSeed(seed);
+//    Knn cls = new Knn();
+    ElasticEnsemble cls = new ElasticEnsemble();
+    cls.setSeed(seed);
     double pp = Double.parseDouble(args[4]);
     System.out.println(pp);
-    ee.setNumParameterSetsPercentage(pp);
-    for(int n = 10; n <= 100; n+=10) { // TODO change!
+    cls.setNumParametersLimitPercentage(pp);
+    for(int n = 10; n <= 100; n+=10) {
         double np = (double) n / 100;
         System.out.println(np + ", " + pp);
         String classifierName = "ee_np=" + np + "_pp=" + pp;
+//        String classifierName = "knn";
         String experimentResultsDirPath = resultsDirPath + "/" + classifierName + "/Predictions/" + datasetName;
         String trainResultsFilePath = experimentResultsDirPath + "/trainFold" + seed + ".csv";
         String testResultsFilePath = experimentResultsDirPath + "/testFold" + seed + ".csv";
-        ee.setNeighbourhoodSizePercentage(np);
+        cls.setNeighbourhoodSizeLimitPercentage(np);
         boolean trainMissing = !exists(trainResultsFilePath);
         boolean testMissing = !exists(testResultsFilePath);
         if(trainMissing || testMissing) {
             System.out.println("training");
-            ee.buildClassifier(train);
+            cls.buildClassifier(train);
         }
         if(trainMissing) {
             System.out.println("getting train results");
-            ClassifierResults trainResults = ee.getTrainResults();
+            ClassifierResults trainResults = cls.getTrainResults();
             trainResults.setDatasetName(datasetName);
             trainResults.setFoldID(seed);
             trainResults.setClassifierName(classifierName);
@@ -69,12 +72,12 @@ public static void main(String[] args) throws
         }
         if(testMissing) {
             System.out.println("getting test results");
-            ClassifierResults testResults = ee.getTestResults(test);
+            ClassifierResults testResults = cls.getTestResults(test);
             testResults.setDatasetName(datasetName);
             testResults.setFoldID(seed);
             testResults.setClassifierName(classifierName);
             writeToFile(testResults, testResultsFilePath);
-            ee.resetTestRandom();
+            cls.resetTestRandom();
         } else {
             System.out.println("test exists");
         }
@@ -162,14 +165,14 @@ public static void main(String[] args) throws
 //        Instances train = data[0];
 //        Instances test = data[1];
 //        ElasticEnsemble ee = new ElasticEnsemble();
-//        ee.setNumParameterSetsPercentage(0.1);
+//        ee.setNumParametersLimitPercentage(0.1);
 //        ee.setSeed(seed);
 //        ee.setNeighbourSearchStrategy(Knn.NeighbourSearchStrategy.RANDOM);
 //        int n = Integer.parseInt(args[4]);
 ////        n *= train.numClasses();
 ////        double pp = Double.parseDouble(args[4]);
 ////        System.out.println(pp);
-////        ee.setNumParameterSetsPercentage(pp);
+////        ee.setNumParametersLimitPercentage(pp);
 ////        for(int n = 5; n <= 30; n+=5) {
 ////            double np = (double) n / 100;
 ////            System.out.println(np + ", " + pp);
@@ -179,8 +182,8 @@ public static void main(String[] args) throws
 //            String experimentResultsDirPath = resultsDirPath + "/" + classifierName + "/Predictions/" + datasetName;
 //            String trainResultsFilePath = experimentResultsDirPath + "/trainFold" + seed + ".csv";
 //            String testResultsFilePath = experimentResultsDirPath + "/testFold" + seed + ".csv";
-//            ee.setNeighbourhoodSize(n);
-////            ee.setNeighbourhoodSizePercentage(np);
+//            ee.setTrainNeighbourhoodSizeLimit(n);
+////            ee.setTrainNeighbourhoodSizeLimitPercentage(np);
 //            boolean trainMissing = !exists(trainResultsFilePath);
 //            boolean testMissing = !exists(testResultsFilePath);
 //            if(trainMissing || testMissing) {

diff --git a/src/main/java/classifiers/distance_based/elastic_ensemble/ElasticEnsemble.java b/src/main/java/classifiers/distance_based/elastic_ensemble/ElasticEnsemble.java
diff --git a/src/main/java/classifiers/distance_based/elastic_ensemble/iteration/DynamicIterator.java b/src/main/java/classifiers/distance_based/elastic_ensemble/iteration/DynamicIterator.java
@@ -1,5 +1,6 @@
 package classifiers.distance_based.elastic_ensemble.iteration;
 
+import java.util.Collection;
 import java.util.Iterator;
 import java.util.ListIterator;
 
@@ -34,4 +35,10 @@ public void set(final A a) {
 
     @Override
     public abstract B iterator();
+
+    public void addAll(Collection<A> collection) {
+        for(A item : collection) {
+            add(item);
+        }
+    }
 }
diff --git a/.../classifiers/distance_based/elastic_ensemble/iteration/linear/AbstractLinearIterator.java b/.../classifiers/distance_based/elastic_ensemble/iteration/linear/AbstractLinearIterator.java
@@ -8,7 +8,7 @@ public abstract class AbstractLinearIterator<A, B extends AbstractLinearIterator
     protected final List<A> values;
     protected int index = 0;
 
-    public AbstractLinearIterator(final Collection<? extends A> values) {
+    public AbstractLinearIterator(final List<A> values) {
         this.values = new ArrayList<>(values);
     }
 
@@ -23,8 +23,8 @@ public AbstractLinearIterator(AbstractLinearIterator<A, B> other) {
 
     @Override
     public void remove() {
-        index--;
         values.remove(index);
+        index--;
         if(index < 0) {
             index = 0;
         }
@@ -37,7 +37,7 @@ public void add(final A a) {
 
     @Override
     public boolean hasNext() {
-        return !values.isEmpty();
+        return index < values.size();
     }
 
     @Override

diff --git a/...ain/java/classifiers/distance_based/elastic_ensemble/iteration/linear/LinearIterator.java b/...ain/java/classifiers/distance_based/elastic_ensemble/iteration/linear/LinearIterator.java
@@ -1,10 +1,11 @@
 package classifiers.distance_based.elastic_ensemble.iteration.linear;
 
 import java.util.Collection;
+import java.util.List;
 
 public class LinearIterator<A> extends AbstractLinearIterator<A, LinearIterator<A>> {
 
-    public LinearIterator(final Collection<? extends A> values) {
+    public LinearIterator(final List<A> values) {
         super(values);
     }
 

diff --git a/.../classifiers/distance_based/elastic_ensemble/iteration/random/AbstractRandomIterator.java b/.../classifiers/distance_based/elastic_ensemble/iteration/random/AbstractRandomIterator.java
@@ -2,35 +2,35 @@
 
 import classifiers.distance_based.elastic_ensemble.iteration.linear.AbstractLinearIterator;
 
+import java.util.ArrayList;
 import java.util.Collection;
+import java.util.List;
 import java.util.Random;
 
 public abstract class AbstractRandomIterator<A, B extends AbstractRandomIterator<A, B>> extends AbstractLinearIterator<A, AbstractRandomIterator<A, B>> {
     protected final Random random;
+    protected long seed;
 
-    public AbstractRandomIterator(final Collection<? extends A> values, final Random random) {
+    public AbstractRandomIterator(final List<A> values, final long seed) {
         super(values);
-        this.random = random;
+        this.random = new Random(seed);
     }
 
     public AbstractRandomIterator(AbstractRandomIterator<A, B> other) {
-        this(other.values, other.random);
+        this(other.values, other.seed);
         index = other.index;
     }
 
-    public AbstractRandomIterator(Random random) {
-        this.random = random;
+    public AbstractRandomIterator(long seed) {
+        this.random = new Random(seed);
     }
 
-    @Override
-    public void remove() {
-        values.remove(index--);
-    }
-
-
     @Override
     public A next() {
-        return values.get(index = random.nextInt(values.size()));
+        index = random.nextInt(values.size());
+        seed = random.nextLong();
+        random.setSeed(seed);
+        return values.get(index);
     }
 
 }
diff --git a/...ssifiers/distance_based/elastic_ensemble/iteration/random/AbstractRoundRobinIterator.java b/...ssifiers/distance_based/elastic_ensemble/iteration/random/AbstractRoundRobinIterator.java
@@ -3,11 +3,12 @@
 import classifiers.distance_based.elastic_ensemble.iteration.linear.AbstractLinearIterator;
 
 import java.util.Collection;
+import java.util.List;
 
 public abstract class AbstractRoundRobinIterator<A, B extends AbstractRoundRobinIterator<A, B>> extends AbstractLinearIterator<A, AbstractRoundRobinIterator<A, B>>
      {
 
-    public AbstractRoundRobinIterator(final Collection<? extends A> values) {
+    public AbstractRoundRobinIterator(final List<A> values) {
         super(values);
     }
 
@@ -20,11 +21,6 @@ public AbstractRoundRobinIterator() {
         super();
     }
 
-    @Override
-    public void remove() {
-        values.remove(index--);
-    }
-
     @Override
     public A next() {
         return values.get(index = (index + 1) % values.size());

diff --git a/...ain/java/classifiers/distance_based/elastic_ensemble/iteration/random/RandomIterator.java b/...ain/java/classifiers/distance_based/elastic_ensemble/iteration/random/RandomIterator.java
@@ -1,25 +1,31 @@
 package classifiers.distance_based.elastic_ensemble.iteration.random;
 
 import java.util.Collection;
+import java.util.List;
 import java.util.Random;
 
 public class RandomIterator<A> extends AbstractRandomIterator<A, RandomIterator<A>> {
 
-    public RandomIterator(final Collection<? extends A> values, final Random random) {
-        super(values, random);
+    public RandomIterator(final List<A> values, final long seed) {
+        super(values, seed);
     }
 
     public RandomIterator(RandomIterator<A> other) {
-        this(other.values, other.random);
+        this(other.values, other.seed);
         index = other.index;
     }
 
-    public RandomIterator(Random random) {
-        super(random);
+    public RandomIterator(long seed) {
+        super(seed);
     }
 
     @Override
     public RandomIterator<A> iterator() {
         return new RandomIterator<>(this);
     }
+
+    @Override
+    public boolean hasNext() {
+        return !values.isEmpty();
+    }
 }
diff --git a/...java/classifiers/distance_based/elastic_ensemble/iteration/random/RoundRobinIterator.java b/...java/classifiers/distance_based/elastic_ensemble/iteration/random/RoundRobinIterator.java
@@ -1,11 +1,12 @@
 package classifiers.distance_based.elastic_ensemble.iteration.random;
 
 import java.util.Collection;
+import java.util.List;
 import java.util.Random;
 
 public class RoundRobinIterator<A> extends AbstractRoundRobinIterator<A, RoundRobinIterator<A>> {
 
-    public RoundRobinIterator(final Collection<? extends A> values) {
+    public RoundRobinIterator(final List<A> values) {
         super(values);
     }