Author: tommaso
Date: Tue Sep 1 12:26:55 2015
New Revision: 1700527
URL: http://svn.apache.org/r1700527
Log:
OPENNLP-713 - slightly enhanced some tests, made Hypothesis unmutable
Modified:
opennlp/sandbox/nlp-utils/src/main/java/org/apache/opennlp/utils/regression/GradientDescentUtils.java
opennlp/sandbox/nlp-utils/src/main/java/org/apache/opennlp/utils/regression/Hypothesis.java
opennlp/sandbox/nlp-utils/src/main/java/org/apache/opennlp/utils/regression/LinearCombinationHypothesis.java
opennlp/sandbox/nlp-utils/src/test/java/org/apache/opennlp/utils/anomalydetection/AnomalyDetectionUtilsTest.java
opennlp/sandbox/nlp-utils/src/test/java/org/apache/opennlp/utils/regression/GradientDescentUtilsTest.java
opennlp/sandbox/nlp-utils/src/test/java/org/apache/opennlp/utils/regression/RegressionModelUtilsTest.java
Modified:
opennlp/sandbox/nlp-utils/src/main/java/org/apache/opennlp/utils/regression/GradientDescentUtils.java
URL:
http://svn.apache.org/viewvc/opennlp/sandbox/nlp-utils/src/main/java/org/apache/opennlp/utils/regression/GradientDescentUtils.java?rev=1700527&r1=1700526&r2=1700527&view=diff
==============================================================================
---
opennlp/sandbox/nlp-utils/src/main/java/org/apache/opennlp/utils/regression/GradientDescentUtils.java
(original)
+++
opennlp/sandbox/nlp-utils/src/main/java/org/apache/opennlp/utils/regression/GradientDescentUtils.java
Tue Sep 1 12:26:55 2015
@@ -33,14 +33,13 @@ public class GradientDescentUtils {
* Calculates batch gradient descent on the give hypothesis, training set
and learning rate alpha.
* The algorithms iteratively adjusts the hypothesis parameters
*
- * @param hypothesis the hypothesis representing the model used
* @param trainingSet the training set used to fit the parameters
* @param alpha the learning rate alpha used to define how big the
descent steps are
*/
- public static void batchGradientDescent(Hypothesis hypothesis, TrainingSet
trainingSet, double alpha) {
+ public static void batchGradientDescent(TrainingSet trainingSet, double
alpha) {
// set initial random weights
double[] parameters =
initializeRandomWeights(trainingSet.iterator().next().getInputs().length);
- hypothesis.updateParameters(parameters);
+ Hypothesis hypothesis = new LinearCombinationHypothesis(parameters);
int iterations = 0;
@@ -52,7 +51,7 @@ public class GradientDescentUtils {
if (newCost > cost) {
throw new RuntimeException("failed to converge at iteration " +
iterations + " with cost going from " + cost + " to " + newCost);
} else if (cost == newCost || newCost < THRESHOLD || iterations >
MAX_ITERATIONS) {
- System.out.println(cost + " with parameters " +
Arrays.toString(parameters));
+ System.out.println(cost + " with parameters " +
Arrays.toString(parameters) + "(" + iterations + " iterations)");
break;
}
@@ -63,7 +62,7 @@ public class GradientDescentUtils {
parameters = RegressionModelUtils.batchLeastMeanSquareUpdate(parameters,
alpha, trainingSet, hypothesis);
// update weights in the hypothesis
- hypothesis.updateParameters(parameters);
+ hypothesis = new LinearCombinationHypothesis(parameters);
iterations++;
}
Modified:
opennlp/sandbox/nlp-utils/src/main/java/org/apache/opennlp/utils/regression/Hypothesis.java
URL:
http://svn.apache.org/viewvc/opennlp/sandbox/nlp-utils/src/main/java/org/apache/opennlp/utils/regression/Hypothesis.java?rev=1700527&r1=1700526&r2=1700527&view=diff
==============================================================================
---
opennlp/sandbox/nlp-utils/src/main/java/org/apache/opennlp/utils/regression/Hypothesis.java
(original)
+++
opennlp/sandbox/nlp-utils/src/main/java/org/apache/opennlp/utils/regression/Hypothesis.java
Tue Sep 1 12:26:55 2015
@@ -31,10 +31,4 @@ public interface Hypothesis {
*/
double calculateOutput(double[] inputs);
- /**
- * update the internal model's parameters.
- *
- * @param parameters an array of <code>double</code> containing the updated
parameters
- */
- void updateParameters(double[] parameters);
}
Modified:
opennlp/sandbox/nlp-utils/src/main/java/org/apache/opennlp/utils/regression/LinearCombinationHypothesis.java
URL:
http://svn.apache.org/viewvc/opennlp/sandbox/nlp-utils/src/main/java/org/apache/opennlp/utils/regression/LinearCombinationHypothesis.java?rev=1700527&r1=1700526&r2=1700527&view=diff
==============================================================================
---
opennlp/sandbox/nlp-utils/src/main/java/org/apache/opennlp/utils/regression/LinearCombinationHypothesis.java
(original)
+++
opennlp/sandbox/nlp-utils/src/main/java/org/apache/opennlp/utils/regression/LinearCombinationHypothesis.java
Tue Sep 1 12:26:55 2015
@@ -22,7 +22,11 @@ package org.apache.opennlp.utils.regress
* Simplest {@link Hypothesis} which just linearly combines inputs with weights
*/
public class LinearCombinationHypothesis implements Hypothesis {
- private double[] weights;
+ private final double[] weights;
+
+ public LinearCombinationHypothesis(double... weights) {
+ this.weights = weights;
+ }
@Override
public double calculateOutput(double[] inputs) {
@@ -33,8 +37,4 @@ public class LinearCombinationHypothesis
return output;
}
- @Override
- public void updateParameters(double[] parameters) {
- weights = parameters;
- }
-}
+ }
Modified:
opennlp/sandbox/nlp-utils/src/test/java/org/apache/opennlp/utils/anomalydetection/AnomalyDetectionUtilsTest.java
URL:
http://svn.apache.org/viewvc/opennlp/sandbox/nlp-utils/src/test/java/org/apache/opennlp/utils/anomalydetection/AnomalyDetectionUtilsTest.java?rev=1700527&r1=1700526&r2=1700527&view=diff
==============================================================================
---
opennlp/sandbox/nlp-utils/src/test/java/org/apache/opennlp/utils/anomalydetection/AnomalyDetectionUtilsTest.java
(original)
+++
opennlp/sandbox/nlp-utils/src/test/java/org/apache/opennlp/utils/anomalydetection/AnomalyDetectionUtilsTest.java
Tue Sep 1 12:26:55 2015
@@ -23,6 +23,7 @@ import org.apache.opennlp.utils.Training
import org.apache.opennlp.utils.TrainingSet;
import org.junit.Test;
+import static org.junit.Assert.assertEquals;
import static org.junit.Assert.assertNotNull;
import static org.junit.Assert.assertTrue;
@@ -39,19 +40,17 @@ public class AnomalyDetectionUtilsTest {
assertNotNull(mus);
double[] sigmas = AnomalyDetectionUtils.fitSigmas(mus, trainingSet);
assertNotNull(sigmas);
- TrainingExample newInput = new TrainingExample(new double[]{1d, 2d, 1000d,
123d, 0.1d}, 0d);
+ TrainingExample newInput = new TrainingExample(new
double[]{0.4d,0.5d,0.5d,0.5d,0.2d}, 0d);
double probability =
AnomalyDetectionUtils.getGaussianProbability(newInput, mus, sigmas);
- assertTrue("negative probability " + probability, 0 <= probability);
- assertTrue("probability bigger than 1 " + probability, 1 >= probability);
+ assertEquals(0.5d, probability, 0.5d);
}
@Test
public void testGaussianDistributionProbabilityFromTrainingSet() throws
Exception {
TrainingSet trainingSet = new TrainingSet();
TestUtils.fillTrainingSet(trainingSet, 100, 5);
- TrainingExample newInput = new TrainingExample(new double[]{1d, 2d, 1000d,
123d, 0.1d}, 0d);
+ TrainingExample newInput = new TrainingExample(new
double[]{0.4d,0.5d,0.5d,0.5d,0.2d}, 0d);
double probability =
AnomalyDetectionUtils.getGaussianProbability(newInput, trainingSet);
- assertTrue("negative probability " + probability, 0 <= probability);
- assertTrue("probability bigger than 1 " + probability, 1 >= probability);
+ assertEquals(0.5d, probability, 0.5d);
}
}
Modified:
opennlp/sandbox/nlp-utils/src/test/java/org/apache/opennlp/utils/regression/GradientDescentUtilsTest.java
URL:
http://svn.apache.org/viewvc/opennlp/sandbox/nlp-utils/src/test/java/org/apache/opennlp/utils/regression/GradientDescentUtilsTest.java?rev=1700527&r1=1700526&r2=1700527&view=diff
==============================================================================
---
opennlp/sandbox/nlp-utils/src/test/java/org/apache/opennlp/utils/regression/GradientDescentUtilsTest.java
(original)
+++
opennlp/sandbox/nlp-utils/src/test/java/org/apache/opennlp/utils/regression/GradientDescentUtilsTest.java
Tue Sep 1 12:26:55 2015
@@ -31,7 +31,7 @@ public class GradientDescentUtilsTest {
public void testConvergence() throws Exception {
TrainingSet trainingSet = new TrainingSet();
TestUtils.fillTrainingSet(trainingSet, 100, 5);
- GradientDescentUtils.batchGradientDescent(new
LinearCombinationHypothesis(), trainingSet, 0.00002);
+ GradientDescentUtils.batchGradientDescent(trainingSet, 0.00002);
}
}
Modified:
opennlp/sandbox/nlp-utils/src/test/java/org/apache/opennlp/utils/regression/RegressionModelUtilsTest.java
URL:
http://svn.apache.org/viewvc/opennlp/sandbox/nlp-utils/src/test/java/org/apache/opennlp/utils/regression/RegressionModelUtilsTest.java?rev=1700527&r1=1700526&r2=1700527&view=diff
==============================================================================
---
opennlp/sandbox/nlp-utils/src/test/java/org/apache/opennlp/utils/regression/RegressionModelUtilsTest.java
(original)
+++
opennlp/sandbox/nlp-utils/src/test/java/org/apache/opennlp/utils/regression/RegressionModelUtilsTest.java
Tue Sep 1 12:26:55 2015
@@ -34,8 +34,7 @@ public class RegressionModelUtilsTest {
public void testLMS() throws Exception {
TrainingSet trainingSet = new TrainingSet();
trainingSet.add(new TrainingExample(new double[]{10, 10}, 1));
- LinearCombinationHypothesis hypothesis = new LinearCombinationHypothesis();
- hypothesis.updateParameters(new double[]{1, 1});
+ LinearCombinationHypothesis hypothesis = new
LinearCombinationHypothesis(1, 1);
double[] updatedParameters =
RegressionModelUtils.batchLeastMeanSquareUpdate(new double[]{1, 1}, 0.1,
trainingSet, hypothesis);
assertNotNull(updatedParameters);
assertTrue(updatedParameters.length == 2);
