joel/largeRCRF-Java
1
0
Fork 0
Code Issues 3 Pull requests Releases Wiki Activity

Add tests around NumericSplitRuleUpdater; fix minor bug.

Browse source
This commit is contained in:
Joel Therrien 2019-01-23 19:37:10 -08:00
parent 115c57f829
commit d8e52ecd82
3 changed files with 204 additions and 4 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
src/main/java/ca/joeltherrien/randomforest/covariates/numeric/NumericSplitRuleUpdater.java
View file

@ -41,7 +41,7 @@ public class NumericSplitRuleUpdater<Y> implements Covariate.SplitRuleUpdater<Y,
final List<Row<Y>> rightHandList = orderedData;
this.currentSplit = new Split<>(
covariate.new NumericSplitRule(Double.MIN_VALUE),
covariate.new NumericSplitRule(Double.NEGATIVE_INFINITY),
leftHandList,
rightHandList,
Collections.emptyList());

6
src/main/java/ca/joeltherrien/randomforest/covariates/numeric/NumericSplitUpdate.java
View file

@ -20,13 +20,13 @@ import ca.joeltherrien.randomforest.Row;
import ca.joeltherrien.randomforest.covariates.Covariate;
import lombok.AllArgsConstructor;
import java.util.Collection;
import java.util.List;
@AllArgsConstructor
public class NumericSplitUpdate<Y> implements Covariate.SplitUpdate<Y, Double> {
private final NumericCovariate.NumericSplitRule numericSplitRule;
private final Collection<Row<Y>> rowsMoved;
private final List<Row<Y>> rowsMoved;
@Override
public NumericCovariate.NumericSplitRule getSplitRule() {
@ -34,7 +34,7 @@ public class NumericSplitUpdate<Y> implements Covariate.SplitUpdate<Y, Double> {
}
@Override
public Collection<Row<Y>> rowsMovedToLeftHand() {
public List<Row<Y>> rowsMovedToLeftHand() {
return rowsMoved;
}
}

200
src/test/java/ca/joeltherrien/randomforest/covariates/NumericCovariateTest.java Normal file
View file

@ -0,0 +1,200 @@
package ca.joeltherrien.randomforest.covariates;
import ca.joeltherrien.randomforest.Row;
import ca.joeltherrien.randomforest.covariates.numeric.NumericCovariate;
import ca.joeltherrien.randomforest.covariates.numeric.NumericSplitRuleUpdater;
import ca.joeltherrien.randomforest.covariates.numeric.NumericSplitUpdate;
import ca.joeltherrien.randomforest.tree.Split;
import ca.joeltherrien.randomforest.utils.IndexedIterator;
import ca.joeltherrien.randomforest.utils.UniqueSubsetValueIterator;
import ca.joeltherrien.randomforest.utils.UniqueValueIterator;
import ca.joeltherrien.randomforest.utils.Utils;
import org.junit.jupiter.api.Test;
import java.util.ArrayList;
import java.util.Collections;
import java.util.List;
import java.util.Random;
import static org.junit.jupiter.api.Assertions.*;
public class NumericCovariateTest {
private List<Row<Double>> createTestDataset(NumericCovariate covariate){
final List<Row<Double>> rowList = new ArrayList<>();
final List<Covariate> covariateList = Collections.singletonList(covariate);
rowList.add(Row.createSimple(Utils.easyMap("x", "1.0"), covariateList, 1, 1.0));
rowList.add(Row.createSimple(Utils.easyMap("x", "2.0"), covariateList, 2, 1.0));
rowList.add(Row.createSimple(Utils.easyMap("x", "2.0"), covariateList, 3, 1.0));
rowList.add(Row.createSimple(Utils.easyMap("x", "3.0"), covariateList, 4, 1.0));
rowList.add(Row.createSimple(Utils.easyMap("x", "3.0"), covariateList, 5, 1.0));
rowList.add(Row.createSimple(Utils.easyMap("x", "4.0"), covariateList, 6, 1.0));
rowList.add(Row.createSimple(Utils.easyMap("x", "4.0"), covariateList, 7, 1.0));
return rowList;
}
@Test
public void testNumericCovariateDeterministic(){
final NumericCovariate covariate = new NumericCovariate("x", 0);
final List<Row<Double>> dataset = createTestDataset(covariate);
// nsplit=0 so result should be deterministic
final NumericSplitRuleUpdater<Double> splitRuleUpdater = covariate.generateSplitRuleUpdater(dataset, 0, new Random());
// Current Split should be empty on the left side
Split<Double, Double> currentSplit = splitRuleUpdater.currentSplit();
assertEquals(0, currentSplit.getLeftHand().size());
assertEquals(dataset.size(), currentSplit.getRightHand().size());
assertFalse(currentSplit.getSplitRule().isLeftHand(covariate.createValue(-1.0))); // everything is right hand; we have no negative values
/* First Update */
assertTrue(splitRuleUpdater.hasNext());
NumericSplitUpdate<Double> update = splitRuleUpdater.nextUpdate();
assertEquals(1, update.rowsMovedToLeftHand().size());
assertEquals(dataset.get(0), update.rowsMovedToLeftHand().get(0));
int index = 1;
currentSplit = splitRuleUpdater.currentSplit();
assertEquals(index, currentSplit.getLeftHand().size());
assertContains(dataset.subList(0,index), currentSplit.getLeftHand());
assertEquals(dataset.size()-index, currentSplit.getRightHand().size());
assertContains(dataset.subList(index, dataset.size()), currentSplit.getRightHand());
assertTrue(currentSplit.getSplitRule().isLeftHand(covariate.createValue(1.0)));
assertFalse(currentSplit.getSplitRule().isLeftHand(covariate.createValue(1.5)));
/* Second Update */
assertTrue(splitRuleUpdater.hasNext());
update = splitRuleUpdater.nextUpdate();
assertEquals(2, update.rowsMovedToLeftHand().size());
assertTrue(update.rowsMovedToLeftHand().contains(dataset.get(1)));
assertTrue(update.rowsMovedToLeftHand().contains(dataset.get(2)));
index = 3;
currentSplit = splitRuleUpdater.currentSplit();
assertEquals(index, currentSplit.getLeftHand().size());
assertContains(dataset.subList(0,index), currentSplit.getLeftHand());
assertEquals(dataset.size()-index, currentSplit.getRightHand().size());
assertContains(dataset.subList(index, dataset.size()), currentSplit.getRightHand());
assertTrue(currentSplit.getSplitRule().isLeftHand(covariate.createValue(2.0)));
assertFalse(currentSplit.getSplitRule().isLeftHand(covariate.createValue(2.5)));
/* Third Update */
assertTrue(splitRuleUpdater.hasNext());
update = splitRuleUpdater.nextUpdate();
assertEquals(2, update.rowsMovedToLeftHand().size());
assertTrue(update.rowsMovedToLeftHand().contains(dataset.get(3)));
assertTrue(update.rowsMovedToLeftHand().contains(dataset.get(4)));
index = 5;
currentSplit = splitRuleUpdater.currentSplit();
assertEquals(index, currentSplit.getLeftHand().size());
assertContains(dataset.subList(0,index), currentSplit.getLeftHand());
assertEquals(dataset.size()-index, currentSplit.getRightHand().size());
assertContains(dataset.subList(index, dataset.size()), currentSplit.getRightHand());
assertTrue(currentSplit.getSplitRule().isLeftHand(covariate.createValue(3.0)));
assertFalse(currentSplit.getSplitRule().isLeftHand(covariate.createValue(3.5)));
/* Fourth Update */
assertTrue(splitRuleUpdater.hasNext());
update = splitRuleUpdater.nextUpdate();
assertEquals(2, update.rowsMovedToLeftHand().size());
assertTrue(update.rowsMovedToLeftHand().contains(dataset.get(5)));
assertTrue(update.rowsMovedToLeftHand().contains(dataset.get(6)));
index = 7;
currentSplit = splitRuleUpdater.currentSplit();
assertEquals(index, currentSplit.getLeftHand().size());
assertContains(dataset.subList(0,index), currentSplit.getLeftHand());
assertEquals(dataset.size()-index, currentSplit.getRightHand().size());
assertContains(dataset.subList(index, dataset.size()), currentSplit.getRightHand());
assertTrue(currentSplit.getSplitRule().isLeftHand(covariate.createValue(4.0)));
assertFalse(currentSplit.getSplitRule().isLeftHand(covariate.createValue(4.5)));
assertFalse(splitRuleUpdater.hasNext());
}
@Test
public void testNumericSplitRuleUpdaterWithIndexes(){
final NumericCovariate covariate = new NumericCovariate("x", 0);
final List<Row<Double>> dataset = createTestDataset(covariate);
final Integer[] indexes = new Integer[]{2,3,4};
final IndexedIterator<Double> dataIterator = new UniqueSubsetValueIterator<>(
new UniqueValueIterator<>(dataset.stream()
.map(row -> row.getCovariateValue(covariate).getValue()).iterator()),
indexes
);
final NumericSplitRuleUpdater<Double> splitRuleUpdater = new NumericSplitRuleUpdater<>(covariate, dataset, dataIterator);
// Current Split should be empty on the left side
Split<Double, Double> currentSplit = splitRuleUpdater.currentSplit();
assertEquals(0, currentSplit.getLeftHand().size());
assertEquals(dataset.size(), currentSplit.getRightHand().size());
assertFalse(currentSplit.getSplitRule().isLeftHand(covariate.createValue(-1.0))); // everything is right hand; we have no negative values
/* First Update */
assertTrue(splitRuleUpdater.hasNext());
NumericSplitUpdate<Double> update = splitRuleUpdater.nextUpdate();
assertEquals(3, update.rowsMovedToLeftHand().size());
assertTrue(update.rowsMovedToLeftHand().contains(dataset.get(0)));
assertTrue(update.rowsMovedToLeftHand().contains(dataset.get(1)));
assertTrue(update.rowsMovedToLeftHand().contains(dataset.get(2)));
int index = 3;
currentSplit = splitRuleUpdater.currentSplit();
assertEquals(index, currentSplit.getLeftHand().size());
assertContains(dataset.subList(0,index), currentSplit.getLeftHand());
assertEquals(dataset.size()-index, currentSplit.getRightHand().size());
assertContains(dataset.subList(index, dataset.size()), currentSplit.getRightHand());
assertTrue(currentSplit.getSplitRule().isLeftHand(covariate.createValue(2.0)));
assertFalse(currentSplit.getSplitRule().isLeftHand(covariate.createValue(2.5)));
/* Second Update */
assertTrue(splitRuleUpdater.hasNext());
update = splitRuleUpdater.nextUpdate();
assertEquals(2, update.rowsMovedToLeftHand().size());
assertTrue(update.rowsMovedToLeftHand().contains(dataset.get(3)));
assertTrue(update.rowsMovedToLeftHand().contains(dataset.get(4)));
index = 5;
currentSplit = splitRuleUpdater.currentSplit();
assertEquals(index, currentSplit.getLeftHand().size());
assertContains(dataset.subList(0,index), currentSplit.getLeftHand());
assertEquals(dataset.size()-index, currentSplit.getRightHand().size());
assertContains(dataset.subList(index, dataset.size()), currentSplit.getRightHand());
assertTrue(currentSplit.getSplitRule().isLeftHand(covariate.createValue(3.0)));
assertFalse(currentSplit.getSplitRule().isLeftHand(covariate.createValue(3.5)));
assertFalse(splitRuleUpdater.hasNext());
}
private <T> void assertContains(List<T> subList, List<T> greaterList){
boolean allContained = true;
for(T candidate : subList){
if(!greaterList.contains(candidate)){
allContained = false;
break;
}
}
assertTrue(allContained);
}
}