This is an automated email from the ASF dual-hosted git repository.
zaleslaw pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/ignite.git
The following commit(s) were added to refs/heads/master by this push:
new 45024f07 IGNITE-9978: [ML] Implement Compound Naive Bayes classifier
(#6567)
45024f07 is described below
commit 45024f07c0ce8c05a0abbba512da57e2cc5be435
Author: Ravil Galeyev <deh...@yandex.ru>
AuthorDate: Sun Sep 8 11:22:34 2019 +0200
IGNITE-9978: [ML] Implement Compound Naive Bayes classifier (#6567)
---
.../ml/naivebayes/CompoundNaiveBayesExample.java | 95 ++++++++++++
.../apache/ignite/ml/naivebayes/BayesModel.java | 34 +++++
.../compound/CompoundNaiveBayesModel.java | 159 ++++++++++++++++++++
.../compound/CompoundNaiveBayesTrainer.java | 160 +++++++++++++++++++++
.../ml/naivebayes/compound/package-info.java | 22 +++
.../discrete/DiscreteNaiveBayesModel.java | 56 +++++---
.../discrete/DiscreteNaiveBayesSumsHolder.java | 2 +-
.../discrete/DiscreteNaiveBayesTrainer.java | 2 +-
.../gaussian/GaussianNaiveBayesModel.java | 55 ++++---
.../gaussian/GaussianNaiveBayesSumsHolder.java | 2 +-
.../gaussian/GaussianNaiveBayesTrainer.java | 2 +-
.../apache/ignite/ml/util/MLSandboxDatasets.java | 5 +-
.../src/main/resources/datasets/mixed_dataset.csv | 8 ++
.../compound/CompoundNaiveBayesModelTest.java | 95 ++++++++++++
.../compound/CompoundNaiveBayesTest.java | 65 +++++++++
.../compound/CompoundNaiveBayesTrainerTest.java | 108 ++++++++++++++
.../apache/ignite/ml/naivebayes/compound/Data.java | 88 ++++++++++++
.../discrete/DiscreteNaiveBayesModelTest.java | 6 +-
.../discrete/DiscreteNaiveBayesTrainerTest.java | 12 +-
.../gaussian/GaussianNaiveBayesModelTest.java | 3 +-
.../gaussian/GaussianNaiveBayesTrainerTest.java | 22 +--
21 files changed, 932 insertions(+), 69 deletions(-)
diff --git
a/examples/src/main/java/org/apache/ignite/examples/ml/naivebayes/CompoundNaiveBayesExample.java
b/examples/src/main/java/org/apache/ignite/examples/ml/naivebayes/CompoundNaiveBayesExample.java
new file mode 100644
index 0000000..5c0aac2
--- /dev/null
+++
b/examples/src/main/java/org/apache/ignite/examples/ml/naivebayes/CompoundNaiveBayesExample.java
@@ -0,0 +1,95 @@
+/*
+ * 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.ignite.examples.ml.naivebayes;
+
+import java.io.FileNotFoundException;
+import org.apache.ignite.Ignite;
+import org.apache.ignite.IgniteCache;
+import org.apache.ignite.Ignition;
+import org.apache.ignite.ml.dataset.feature.extractor.Vectorizer;
+import org.apache.ignite.ml.dataset.feature.extractor.impl.DummyVectorizer;
+import org.apache.ignite.ml.math.primitives.vector.Vector;
+import org.apache.ignite.ml.naivebayes.compound.CompoundNaiveBayesModel;
+import org.apache.ignite.ml.naivebayes.compound.CompoundNaiveBayesTrainer;
+import org.apache.ignite.ml.naivebayes.discrete.DiscreteNaiveBayesTrainer;
+import org.apache.ignite.ml.naivebayes.gaussian.GaussianNaiveBayesTrainer;
+import org.apache.ignite.ml.selection.scoring.evaluator.Evaluator;
+import org.apache.ignite.ml.util.MLSandboxDatasets;
+import org.apache.ignite.ml.util.SandboxMLCache;
+
+import static java.util.Arrays.asList;
+
+/**
+ * Run naive Compound Bayes classification model based on <a
href="https://en.wikipedia.org/wiki/Naive_Bayes_classifier";>
+ * Nnaive Bayes classifier</a> algorithm ({@link
GaussianNaiveBayesTrainer})and <a
+ *
href=https://en.wikipedia.org/wiki/Naive_Bayes_classifier#Multinomial_naive_Bayes";>
Discrete naive Bayes
+ * classifier</a> algorithm ({@link DiscreteNaiveBayesTrainer}) over
distributed cache.
+ * <p>
+ * Code in this example launches Ignite grid and fills the cache with test
data points.
+ * <p>
+ * After that it trains the naive Bayes classification model based on the
specified data.</p>
+ * <p>
+ * Finally, this example loops over the test set of data points, applies the
trained model to predict the target value,
+ * compares prediction to expected outcome (ground truth), and builds
+ * <a href="https://en.wikipedia.org/wiki/Confusion_matrix";>confusion
matrix</a>.</p>
+ * <p>
+ * You can change the test data used in this example and re-run it to explore
this algorithm further.</p>
+ */
+public class CompoundNaiveBayesExample {
+ public static void main(String[] args) throws FileNotFoundException {
+ System.out.println();
+ System.out.println(">>> Compound Naive Bayes classification model over
partitioned dataset usage example started.");
+ // Start ignite grid.
+ try (Ignite ignite =
Ignition.start("examples/config/example-ignite.xml")) {
+ System.out.println(">>> Ignite grid started.");
+
+ IgniteCache<Integer, Vector> dataCache = new SandboxMLCache(ignite)
+ .fillCacheWith(MLSandboxDatasets.MIXED_DATASET);
+
+ double[] priorProbabilities = new double[] {.5, .5};
+ double[][] thresholds = new double[][] {{.5}, {.5}, {.5}, {.5},
{.5}};
+
+ System.out.println(">>> Create new naive Bayes classification
trainer object.");
+ CompoundNaiveBayesTrainer trainer = new CompoundNaiveBayesTrainer()
+ .withPriorProbabilities(priorProbabilities)
+ .withGaussianNaiveBayesTrainer(new GaussianNaiveBayesTrainer())
+ .withGaussianFeatureIdsToSkip(asList(3, 4, 5, 6, 7))
+ .withDiscreteNaiveBayesTrainer(new DiscreteNaiveBayesTrainer()
+ .setBucketThresholds(thresholds))
+ .withDiscreteFeatureIdsToSkip(asList(0, 1, 2));
+ System.out.println(">>> Perform the training to get the model.");
+
+ Vectorizer<Integer, Vector, Integer, Double> vectorizer = new
DummyVectorizer<Integer>()
+ .labeled(Vectorizer.LabelCoordinate.FIRST);
+
+ CompoundNaiveBayesModel mdl = trainer.fit(ignite, dataCache,
vectorizer);
+
+ System.out.println(">>> Compound Naive Bayes model: " + mdl);
+
+ double accuracy = Evaluator.evaluate(
+ dataCache,
+ mdl,
+ vectorizer
+ ).accuracy();
+
+ System.out.println("\n>>> Accuracy " + accuracy);
+
+ System.out.println(">>> Compound Naive bayes model over
partitioned dataset usage example completed.");
+ }
+ }
+}
diff --git
a/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/BayesModel.java
b/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/BayesModel.java
new file mode 100644
index 0000000..d08b720
--- /dev/null
+++ b/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/BayesModel.java
@@ -0,0 +1,34 @@
+/*
+ * 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.ignite.ml.naivebayes;
+
+import org.apache.ignite.ml.Exportable;
+import org.apache.ignite.ml.IgniteModel;
+
+/**
+ * Interface for Bayes Models.
+ */
+public interface BayesModel<MODEL extends BayesModel, FEATURES, OUTPUT>
+ extends IgniteModel<FEATURES, OUTPUT>, Exportable<MODEL> {
+
+ /**
+ * Returns an array where the index correapons a label, and value
corresponds {@code log(probalility)} to be this label.
+ * The prior probabilities are not count.
+ */
+ double[] probabilityPowers(FEATURES vector);
+}
diff --git
a/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/compound/CompoundNaiveBayesModel.java
b/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/compound/CompoundNaiveBayesModel.java
new file mode 100644
index 0000000..5daebb7
--- /dev/null
+++
b/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/compound/CompoundNaiveBayesModel.java
@@ -0,0 +1,159 @@
+/*
+ * 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.ignite.ml.naivebayes.compound;
+
+import java.io.Serializable;
+import java.util.Collection;
+import java.util.Collections;
+
+import org.apache.ignite.ml.Exportable;
+import org.apache.ignite.ml.Exporter;
+import org.apache.ignite.ml.IgniteModel;
+import org.apache.ignite.ml.math.primitives.vector.Vector;
+import org.apache.ignite.ml.math.primitives.vector.VectorUtils;
+import org.apache.ignite.ml.naivebayes.discrete.DiscreteNaiveBayesModel;
+import org.apache.ignite.ml.naivebayes.gaussian.GaussianNaiveBayesModel;
+
+/**
+ * A compound Naive Bayes model which uses a composition of{@code
GaussianNaiveBayesModel} and {@code
+ * DiscreteNaiveBayesModel}.
+ */
+public class CompoundNaiveBayesModel implements IgniteModel<Vector, Double>,
Exportable<CompoundNaiveBayesModel>, Serializable {
+ /** Serial version uid. */
+ private static final long serialVersionUID = -5045925321135798960L;
+
+ /** Prior probabilities of each class. */
+ private double[] priorProbabilities;
+
+ /** Labels. */
+ private double[] labels;
+
+ /** Gaussian Bayes model. */
+ private GaussianNaiveBayesModel gaussianModel;
+
+ /** Feature ids which should be skipped in Gaussian model. */
+ private Collection<Integer> gaussianFeatureIdsToSkip =
Collections.emptyList();
+
+ /** Discrete Bayes model. */
+ private DiscreteNaiveBayesModel discreteModel;
+
+ /** Feature ids which should be skipped in Discrete model. */
+ private Collection<Integer> discreteFeatureIdsToSkip =
Collections.emptyList();
+
+ /** {@inheritDoc} */
+ @Override public <P> void saveModel(Exporter<CompoundNaiveBayesModel, P>
exporter, P path) {
+ exporter.save(this, path);
+ }
+
+ /** {@inheritDoc} */
+ @Override public Double predict(Vector vector) {
+ double[] probapilityPowers = new double[priorProbabilities.length];
+ for (int i = 0; i < priorProbabilities.length; i++) {
+ probapilityPowers[i] = Math.log(priorProbabilities[i]);
+ }
+
+ if (discreteModel != null) {
+ probapilityPowers = sum(probapilityPowers,
discreteModel.probabilityPowers(skipFeatures(vector,
discreteFeatureIdsToSkip)));
+ }
+
+ if (gaussianModel != null) {
+ probapilityPowers = sum(probapilityPowers,
gaussianModel.probabilityPowers(skipFeatures(vector,
gaussianFeatureIdsToSkip)));
+ }
+
+ int maxLabelIndex = 0;
+ for (int i = 0; i < probapilityPowers.length; i++) {
+ if (probapilityPowers[i] > probapilityPowers[maxLabelIndex]) {
+ maxLabelIndex = i;
+ }
+ }
+ return labels[maxLabelIndex];
+ }
+
+ /** Returns a gaussian model. */
+ public GaussianNaiveBayesModel getGaussianModel() {
+ return gaussianModel;
+ }
+
+ /** Returns a discrete model. */
+ public DiscreteNaiveBayesModel getDiscreteModel() {
+ return discreteModel;
+ }
+
+ /** Sets prior probabilities. */
+ public CompoundNaiveBayesModel wirhPriorProbabilities(double[]
priorProbabilities) {
+ this.priorProbabilities = priorProbabilities.clone();
+ return this;
+ }
+
+ /** Sets labels. */
+ public CompoundNaiveBayesModel withLabels(double[] labels) {
+ this.labels = labels.clone();
+ return this;
+ }
+
+ /** Sets a gaussian model. */
+ public CompoundNaiveBayesModel withGaussianModel(GaussianNaiveBayesModel
gaussianModel) {
+ this.gaussianModel = gaussianModel;
+ return this;
+ }
+
+ /** Sets a discrete model. */
+ public CompoundNaiveBayesModel withDiscreteModel(DiscreteNaiveBayesModel
discreteModel) {
+ this.discreteModel = discreteModel;
+ return this;
+ }
+
+ /** Sets feature ids to skip in Gaussian Bayes. */
+ public CompoundNaiveBayesModel
withGaussianFeatureIdsToSkip(Collection<Integer> gaussianFeatureIdsToSkip) {
+ this.gaussianFeatureIdsToSkip = gaussianFeatureIdsToSkip;
+ return this;
+ }
+
+ /** Sets feature ids to skip in discrete Bayes. */
+ public CompoundNaiveBayesModel
withDiscreteFeatureIdsToSkip(Collection<Integer> discreteFeatureIdsToSkip) {
+ this.discreteFeatureIdsToSkip = discreteFeatureIdsToSkip;
+ return this;
+ }
+
+ /** Returns index by index sum of two arrays. */
+ private static double[] sum(double[] arr1, double[] arr2) {
+ assert arr1.length == arr2.length;
+
+ double[] result = new double[arr1.length];
+
+ for (int i = 0; i < arr1.length; i++) {
+ result[i] = arr1[i] + arr2[i];
+ }
+ return result;
+ }
+
+ /** Returns a new (shorter) vector without features provided in {@param
featureIdsToSkip}. */
+ private static Vector skipFeatures(Vector vector, Collection<Integer>
featureIdsToSkip) {
+ int newSize = vector.size() - featureIdsToSkip.size();
+ double[] newFeaturesValues = new double[newSize];
+
+ int index = 0;
+ for (int j = 0; j < vector.size(); j++) {
+ if(featureIdsToSkip.contains(j)) continue;
+
+ newFeaturesValues[index] = vector.get(j);
+ ++index;
+ }
+ return VectorUtils.of(newFeaturesValues);
+ }
+}
diff --git
a/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/compound/CompoundNaiveBayesTrainer.java
b/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/compound/CompoundNaiveBayesTrainer.java
new file mode 100644
index 0000000..9b8cd46
--- /dev/null
+++
b/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/compound/CompoundNaiveBayesTrainer.java
@@ -0,0 +1,160 @@
+/*
+ * 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.ignite.ml.naivebayes.compound;
+
+import org.apache.ignite.ml.dataset.DatasetBuilder;
+import org.apache.ignite.ml.environment.LearningEnvironmentBuilder;
+import org.apache.ignite.ml.math.functions.IgniteFunction;
+import org.apache.ignite.ml.math.primitives.vector.VectorUtils;
+import org.apache.ignite.ml.naivebayes.discrete.DiscreteNaiveBayesModel;
+import org.apache.ignite.ml.naivebayes.discrete.DiscreteNaiveBayesTrainer;
+import org.apache.ignite.ml.naivebayes.gaussian.GaussianNaiveBayesModel;
+import org.apache.ignite.ml.naivebayes.gaussian.GaussianNaiveBayesTrainer;
+import org.apache.ignite.ml.preprocessing.Preprocessor;
+import org.apache.ignite.ml.structures.LabeledVector;
+import org.apache.ignite.ml.trainers.SingleLabelDatasetTrainer;
+
+import java.util.Collection;
+import java.util.Collections;
+
+/**
+ * Trainer for the compound Naive Bayes classifier model. It uses a model
composition of {@code
+ * GaussianNaiveBayesTrainer} and {@code DiscreteNaiveBayesTrainer}. To
distinguish which features with which trainer
+ * should be used, each trainer should have a collection of feature ids which
should be skipped. It can be set by {@code
+ * #setFeatureIdsToSkip()} method.
+ */
+public class CompoundNaiveBayesTrainer extends
SingleLabelDatasetTrainer<CompoundNaiveBayesModel> {
+
+ /** Prior probabilities of each class. */
+ private double[] priorProbabilities;
+
+ /** Gaussian Naive Bayes trainer. */
+ private GaussianNaiveBayesTrainer gaussianNaiveBayesTrainer;
+
+ /** Feature ids which should be skipped in Gaussian model. */
+ private Collection<Integer> gaussianFeatureIdsToSkip =
Collections.emptyList();
+
+ /** Discrete Naive Bayes trainer. */
+ private DiscreteNaiveBayesTrainer discreteNaiveBayesTrainer;
+
+ /** Feature ids which should be skipped in Discrete model. */
+ private Collection<Integer> discreteFeatureIdsToSkip =
Collections.emptyList();
+
+ /** {@inheritDoc} */
+ @Override public <K, V> CompoundNaiveBayesModel fit(DatasetBuilder<K, V>
datasetBuilder,
+ Preprocessor<K, V> extractor) {
+ return updateModel(null, datasetBuilder, extractor);
+ }
+
+ /** {@inheritDoc} */
+ @Override public boolean isUpdateable(CompoundNaiveBayesModel mdl) {
+ return gaussianNaiveBayesTrainer.isUpdateable(mdl.getGaussianModel())
+ && discreteNaiveBayesTrainer.isUpdateable(mdl.getDiscreteModel());
+ }
+
+ /** {@inheritDoc} */
+ @Override public CompoundNaiveBayesTrainer
withEnvironmentBuilder(LearningEnvironmentBuilder envBuilder) {
+ return
(CompoundNaiveBayesTrainer)super.withEnvironmentBuilder(envBuilder);
+ }
+
+ /** {@inheritDoc} */
+ @Override protected <K, V> CompoundNaiveBayesModel
updateModel(CompoundNaiveBayesModel mdl,
+ DatasetBuilder<K, V> datasetBuilder, Preprocessor<K, V> extractor) {
+
+ CompoundNaiveBayesModel compoundModel = new CompoundNaiveBayesModel()
+ .wirhPriorProbabilities(priorProbabilities);
+
+ if (gaussianNaiveBayesTrainer != null) {
+ if (priorProbabilities != null) {
+
gaussianNaiveBayesTrainer.setPriorProbabilities(priorProbabilities);
+ }
+ GaussianNaiveBayesModel model = (mdl == null)
+ ? gaussianNaiveBayesTrainer.fit(datasetBuilder,
extractor.map(skipFeatures(gaussianFeatureIdsToSkip)))
+ : gaussianNaiveBayesTrainer.update(mdl.getGaussianModel(),
datasetBuilder, extractor.map(skipFeatures(gaussianFeatureIdsToSkip)));
+
+ compoundModel.withGaussianModel(model)
+ .withGaussianFeatureIdsToSkip(gaussianFeatureIdsToSkip)
+ .withLabels(model.getLabels())
+ .wirhPriorProbabilities(priorProbabilities);
+ }
+
+ if (discreteNaiveBayesTrainer != null) {
+ if (priorProbabilities != null) {
+
discreteNaiveBayesTrainer.setPriorProbabilities(priorProbabilities);
+ }
+ DiscreteNaiveBayesModel model = (mdl == null)
+ ? discreteNaiveBayesTrainer.fit(datasetBuilder,
extractor.map(skipFeatures(discreteFeatureIdsToSkip)))
+ : discreteNaiveBayesTrainer.update(mdl.getDiscreteModel(),
datasetBuilder, extractor.map(skipFeatures(discreteFeatureIdsToSkip)));
+
+ compoundModel.withDiscreteModel(model)
+ .withDiscreteFeatureIdsToSkip(discreteFeatureIdsToSkip)
+ .withLabels(model.getLabels())
+ .wirhPriorProbabilities(priorProbabilities);
+ }
+
+ return compoundModel;
+ }
+
+ /** Sets prior probabilities. */
+ public CompoundNaiveBayesTrainer withPriorProbabilities(double[]
priorProbabilities) {
+ this.priorProbabilities = priorProbabilities.clone();
+ return this;
+ }
+
+ /** Sets a gaussian trainer. */
+ public CompoundNaiveBayesTrainer
withGaussianNaiveBayesTrainer(GaussianNaiveBayesTrainer
gaussianNaiveBayesTrainer) {
+ this.gaussianNaiveBayesTrainer = gaussianNaiveBayesTrainer;
+ return this;
+ }
+
+ /** Sets a discrete trainer. */
+ public CompoundNaiveBayesTrainer
withDiscreteNaiveBayesTrainer(DiscreteNaiveBayesTrainer
discreteNaiveBayesTrainer) {
+ this.discreteNaiveBayesTrainer = discreteNaiveBayesTrainer;
+ return this;
+ }
+
+ /** Sets feature ids to skip in Gaussian Bayes. */
+ public CompoundNaiveBayesTrainer
withGaussianFeatureIdsToSkip(Collection<Integer> gaussianFeatureIdsToSkip) {
+ this.gaussianFeatureIdsToSkip = gaussianFeatureIdsToSkip;
+ return this;
+ }
+
+ /** Sets feature ids to skip in discrete Bayes. */
+ public CompoundNaiveBayesTrainer
withDiscreteFeatureIdsToSkip(Collection<Integer> discreteFeatureIdsToSkip) {
+ this.discreteFeatureIdsToSkip = discreteFeatureIdsToSkip;
+ return this;
+ }
+
+ /** Removes features provided in {@param featureIdsToSkip} from a vector.
*/
+ private static IgniteFunction<LabeledVector<Object>,
LabeledVector<Object>> skipFeatures(Collection<Integer> featureIdsToSkip) {
+ return featureValues -> {
+ final int size = featureValues.features().size();
+ int newSize = size - featureIdsToSkip.size();
+
+ double[] newFeaturesValues = new double[newSize];
+ int index = 0;
+ for (int j = 0; j < size; j++) {
+ if(featureIdsToSkip.contains(j)) continue;
+
+ newFeaturesValues[index] = featureValues.get(j);
+ ++index;
+ }
+ return new LabeledVector<>(VectorUtils.of(newFeaturesValues),
featureValues.label());
+ };
+ }
+}
diff --git
a/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/compound/package-info.java
b/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/compound/package-info.java
new file mode 100644
index 0000000..0806c48
--- /dev/null
+++
b/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/compound/package-info.java
@@ -0,0 +1,22 @@
+/*
+ * 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 description. -->
+ * Contains Compound naive Bayes classifier.
+ */
+package org.apache.ignite.ml.naivebayes.compound;
diff --git
a/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/discrete/DiscreteNaiveBayesModel.java
b/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/discrete/DiscreteNaiveBayesModel.java
index 3c35841..44d0767 100644
---
a/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/discrete/DiscreteNaiveBayesModel.java
+++
b/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/discrete/DiscreteNaiveBayesModel.java
@@ -17,20 +17,17 @@
package org.apache.ignite.ml.naivebayes.discrete;
-import org.apache.ignite.ml.Exportable;
import org.apache.ignite.ml.Exporter;
-import org.apache.ignite.ml.IgniteModel;
import org.apache.ignite.ml.math.primitives.vector.Vector;
-
-import java.io.Serializable;
+import org.apache.ignite.ml.naivebayes.BayesModel;
/**
* Discrete naive Bayes model which predicts result value {@code y} belongs to
a class {@code C_k, k in [0..K]} as
* {@code p(C_k,y) =x_1*p_k1^x *...*x_i*p_ki^x_i}. Where {@code x_i} is a
discrete feature, {@code p_ki} is a prior
* probability probability of class {@code p(x|C_k)}. Returns the number of
the most possible class.
*/
-public class DiscreteNaiveBayesModel implements IgniteModel<Vector, Double>,
Exportable<DiscreteNaiveBayesModel>, Serializable {
- /** */
+public class DiscreteNaiveBayesModel implements
BayesModel<DiscreteNaiveBayesModel, Vector, Double> {
+ /** Serial version uid. */
private static final long serialVersionUID = -127386523291350345L;
/**
@@ -58,9 +55,9 @@ public class DiscreteNaiveBayesModel implements
IgniteModel<Vector, Double>, Exp
/**
* @param probabilities Probabilities of features for classes.
* @param clsProbabilities Prior probabilities for classes.
+ * @param labels Labels.
* @param bucketThresholds The threshold to convert a feature to a binary
value.
* @param sumsHolder Amount values which are abouve the threshold per
label.
- * @param labels Labels.
*/
public DiscreteNaiveBayesModel(double[][][] probabilities, double[]
clsProbabilities, double[] labels,
double[][] bucketThresholds, DiscreteNaiveBayesSumsHolder sumsHolder) {
@@ -81,42 +78,57 @@ public class DiscreteNaiveBayesModel implements
IgniteModel<Vector, Double>, Exp
* @return A label with max probability.
*/
@Override public Double predict(Vector vector) {
- double maxProbapilityPower = -Double.MAX_VALUE;
- int maxLabelIndex = -1;
+ double[] probapilityPowers = probabilityPowers(vector);
- for (int i = 0; i < clsProbabilities.length; i++) {
- double probabilityPower = Math.log(clsProbabilities[i]);
+ int maxLabelIndex = 0;
+ for (int i = 0; i < probapilityPowers.length; i++) {
+ probapilityPowers[i] += Math.log(clsProbabilities[i]);
+
+ if (probapilityPowers[i] > probapilityPowers[maxLabelIndex]) {
+ maxLabelIndex = i;
+ }
+ }
+
+ return labels[maxLabelIndex];
+ }
+ /** {@inheritDoc} */
+ @Override
+ public double[] probabilityPowers(Vector vector) {
+ double[] probapilityPowers = new double[clsProbabilities.length];
+
+ for (int i = 0; i < clsProbabilities.length; i++) {
for (int j = 0; j < probabilities[0].length; j++) {
int x = toBucketNumber(vector.get(j), bucketThresholds[j]);
double p = probabilities[i][j][x];
- probabilityPower += (p > 0 ? Math.log(p) : .0);
- }
-
- if (probabilityPower > maxProbapilityPower) {
- maxLabelIndex = i;
- maxProbapilityPower = probabilityPower;
+ probapilityPowers[i] += (p > 0 ? Math.log(p) : .0);
}
}
- return labels[maxLabelIndex];
+
+ return probapilityPowers;
}
- /** */
+ /** A getter for probabilities.*/
public double[][][] getProbabilities() {
return probabilities;
}
- /** */
+ /** A getter for clsProbabilities.*/
public double[] getClsProbabilities() {
return clsProbabilities;
}
- /** */
+ /** A getter for bucketThresholds.*/
public double[][] getBucketThresholds() {
return bucketThresholds;
}
- /** */
+ /** A getter for labels.*/
+ public double[] getLabels() {
+ return labels;
+ }
+
+ /** A getter for sumsHolder.*/
public DiscreteNaiveBayesSumsHolder getSumsHolder() {
return sumsHolder;
}
diff --git
a/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/discrete/DiscreteNaiveBayesSumsHolder.java
b/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/discrete/DiscreteNaiveBayesSumsHolder.java
index 9ea18ca..50b335e 100644
---
a/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/discrete/DiscreteNaiveBayesSumsHolder.java
+++
b/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/discrete/DiscreteNaiveBayesSumsHolder.java
@@ -49,7 +49,7 @@ public class DiscreteNaiveBayesSumsHolder implements
AutoCloseable, Serializable
return arr1;
}
- /** */
+ /** {@inheritDoc} */
@Override public void close() {
// Do nothing, GC will clean up.
}
diff --git
a/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/discrete/DiscreteNaiveBayesTrainer.java
b/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/discrete/DiscreteNaiveBayesTrainer.java
index f5dd6e6..a80c402 100644
---
a/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/discrete/DiscreteNaiveBayesTrainer.java
+++
b/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/discrete/DiscreteNaiveBayesTrainer.java
@@ -115,7 +115,7 @@ public class DiscreteNaiveBayesTrainer extends
SingleLabelDatasetTrainer<Discret
})) {
DiscreteNaiveBayesSumsHolder sumsHolder = dataset.compute(t -> t,
(a, b) -> {
if (a == null)
- return b == null ? new DiscreteNaiveBayesSumsHolder() : b;
+ return b;
if (b == null)
return a;
return a.merge(b);
diff --git
a/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/gaussian/GaussianNaiveBayesModel.java
b/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/gaussian/GaussianNaiveBayesModel.java
index bef2e39..29b8cc2 100644
---
a/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/gaussian/GaussianNaiveBayesModel.java
+++
b/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/gaussian/GaussianNaiveBayesModel.java
@@ -17,18 +17,16 @@
package org.apache.ignite.ml.naivebayes.gaussian;
-import java.io.Serializable;
-import org.apache.ignite.ml.Exportable;
import org.apache.ignite.ml.Exporter;
-import org.apache.ignite.ml.IgniteModel;
import org.apache.ignite.ml.math.primitives.vector.Vector;
+import org.apache.ignite.ml.naivebayes.BayesModel;
/**
* Simple naive Bayes model which predicts result value {@code y} belongs to a
class {@code C_k, k in [0..K]} as {@code
* p(C_k,y) = p(C_k)*p(y_1,C_k) *...*p(y_n,C_k) / p(y)}. Return the number of
the most possible class.
*/
-public class GaussianNaiveBayesModel implements IgniteModel<Vector, Double>,
Exportable<GaussianNaiveBayesModel>, Serializable {
- /** */
+public class GaussianNaiveBayesModel implements
BayesModel<GaussianNaiveBayesModel, Vector, Double> {
+ /** Serial version uid. */
private static final long serialVersionUID = -127386523291350345L;
/** Means of features for all classes. kth row contains means for
labels[k] class. */
@@ -69,48 +67,61 @@ public class GaussianNaiveBayesModel implements
IgniteModel<Vector, Double>, Exp
/** Returns a number of class to which the input belongs. */
@Override public Double predict(Vector vector) {
- int k = classProbabilities.length;
+ double[] probapilityPowers = probabilityPowers(vector);
- double maxProbability = .0;
int max = 0;
+ for (int i = 0; i < probapilityPowers.length; i++) {
+ probapilityPowers[i] += Math.log(classProbabilities[i]);
- for (int i = 0; i < k; i++) {
- double p = classProbabilities[i];
- for (int j = 0; j < vector.size(); j++) {
- double x = vector.get(j);
- double g = gauss(x, means[i][j], variances[i][j]);
- p *= g;
- }
- if (p > maxProbability) {
+ if (probapilityPowers[i] > probapilityPowers[max]) {
max = i;
- maxProbability = p;
}
}
return labels[max];
}
- /** */
+ /** {@inheritDoc} */
+ @Override
+ public double[] probabilityPowers(Vector vector) {
+ double[] probapilityPowers = new double[classProbabilities.length];
+
+ for (int i = 0; i < classProbabilities.length; i++) {
+ for (int j = 0; j < vector.size(); j++) {
+ double x = vector.get(j);
+ double parobability = gauss(x, means[i][j], variances[i][j]);
+ probapilityPowers[i] += (parobability > 0 ?
Math.log(parobability) : .0);
+ }
+ }
+ return probapilityPowers;
+ }
+
+ /** A getter for means.*/
public double[][] getMeans() {
return means;
}
- /** */
+ /** A getter for variances.*/
public double[][] getVariances() {
return variances;
}
- /** */
+ /** A getter for classProbabilities.*/
public double[] getClassProbabilities() {
return classProbabilities;
}
- /** */
+ /** A getter for labels.*/
+ public double[] getLabels() {
+ return labels;
+ }
+
+ /** A getter for sumsHolder.*/
public GaussianNaiveBayesSumsHolder getSumsHolder() {
return sumsHolder;
}
- /** Gauss distribution */
- private double gauss(double x, double mean, double variance) {
+ /** Gauss distribution. */
+ private static double gauss(double x, double mean, double variance) {
return Math.exp(-1. * Math.pow(x - mean, 2) / (2. * variance)) /
Math.sqrt(2. * Math.PI * variance);
}
}
diff --git
a/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/gaussian/GaussianNaiveBayesSumsHolder.java
b/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/gaussian/GaussianNaiveBayesSumsHolder.java
index eecec74..7b95ff8 100644
---
a/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/gaussian/GaussianNaiveBayesSumsHolder.java
+++
b/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/gaussian/GaussianNaiveBayesSumsHolder.java
@@ -51,7 +51,7 @@ class GaussianNaiveBayesSumsHolder implements Serializable,
AutoCloseable {
return arr1;
}
- /** */
+ /** {@inheritDoc} */
@Override public void close() {
// Do nothing, GC will clean up.
}
diff --git
a/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/gaussian/GaussianNaiveBayesTrainer.java
b/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/gaussian/GaussianNaiveBayesTrainer.java
index 40ca840..84a16ad 100644
---
a/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/gaussian/GaussianNaiveBayesTrainer.java
+++
b/modules/ml/src/main/java/org/apache/ignite/ml/naivebayes/gaussian/GaussianNaiveBayesTrainer.java
@@ -106,7 +106,7 @@ public class GaussianNaiveBayesTrainer extends
SingleLabelDatasetTrainer<Gaussia
)) {
GaussianNaiveBayesSumsHolder sumsHolder = dataset.compute(t -> t,
(a, b) -> {
if (a == null)
- return b == null ? new GaussianNaiveBayesSumsHolder() : b;
+ return b;
if (b == null)
return a;
return a.merge(b);
diff --git
a/modules/ml/src/main/java/org/apache/ignite/ml/util/MLSandboxDatasets.java
b/modules/ml/src/main/java/org/apache/ignite/ml/util/MLSandboxDatasets.java
index 28e4c9a..6574902 100644
--- a/modules/ml/src/main/java/org/apache/ignite/ml/util/MLSandboxDatasets.java
+++ b/modules/ml/src/main/java/org/apache/ignite/ml/util/MLSandboxDatasets.java
@@ -59,7 +59,10 @@ public enum MLSandboxDatasets {
WHOLESALE_CUSTOMERS("modules/ml/src/main/resources/datasets/wholesale_customers.csv",
true, ","),
/** Fraud detection problem [part of whole dataset]. Could be found <a
href="https://www.kaggle.com/mlg-ulb/creditcardfraud/";>here</a>. */
-
FRAUD_DETECTION("modules/ml/src/main/resources/datasets/fraud_detection.csv",
false, ",");
+
FRAUD_DETECTION("modules/ml/src/main/resources/datasets/fraud_detection.csv",
false, ","),
+
+ /** A dataset with discrete and continious features. */
+ MIXED_DATASET("modules/ml/src/main/resources/datasets/mixed_dataset.csv",
true, ",");
/** Filename. */
private final String filename;
diff --git a/modules/ml/src/main/resources/datasets/mixed_dataset.csv
b/modules/ml/src/main/resources/datasets/mixed_dataset.csv
new file mode 100644
index 0000000..5a0df2b
--- /dev/null
+++ b/modules/ml/src/main/resources/datasets/mixed_dataset.csv
@@ -0,0 +1,8 @@
+1, 6, 180, 12, 0, 0, 1, 1, 1
+1, 5.92, 190, 11, 1, 0, 1, 1, 0
+1, 5.58, 170, 12, 1, 1, 0, 0, 1
+1, 5.92, 165, 10, 1, 1, 0, 0, 0
+0, 5, 100, 6, 1, 0, 0, 1, 1
+0, 5.5, 150, 8, 1, 1, 0, 0, 1
+0, 5.42, 130, 7, 1, 1, 1, 1, 0
+0, 5.75, 150, 9, 1, 1, 0, 1, 0
\ No newline at end of file
diff --git
a/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/compound/CompoundNaiveBayesModelTest.java
b/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/compound/CompoundNaiveBayesModelTest.java
new file mode 100644
index 0000000..227a179
--- /dev/null
+++
b/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/compound/CompoundNaiveBayesModelTest.java
@@ -0,0 +1,95 @@
+/*
+ * 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.ignite.ml.naivebayes.compound;
+
+import org.apache.ignite.ml.math.primitives.vector.Vector;
+import org.apache.ignite.ml.math.primitives.vector.VectorUtils;
+import org.apache.ignite.ml.naivebayes.discrete.DiscreteNaiveBayesModel;
+import org.apache.ignite.ml.naivebayes.gaussian.GaussianNaiveBayesModel;
+import org.junit.Test;
+
+import static java.util.Arrays.asList;
+import static org.apache.ignite.ml.naivebayes.compound.Data.LABEL_2;
+import static
org.apache.ignite.ml.naivebayes.compound.Data.binarizedDataThresholds;
+import static org.apache.ignite.ml.naivebayes.compound.Data.classProbabilities;
+import static org.apache.ignite.ml.naivebayes.compound.Data.labels;
+import static org.apache.ignite.ml.naivebayes.compound.Data.means;
+import static org.apache.ignite.ml.naivebayes.compound.Data.probabilities;
+import static org.apache.ignite.ml.naivebayes.compound.Data.variances;
+import static org.junit.Assert.assertEquals;
+
+/** Tests for {@link CompoundNaiveBayesModel} */
+public class CompoundNaiveBayesModelTest {
+
+ /** Precision in test checks. */
+ private static final double PRECISION = 1e-2;
+
+ /** Test. */
+ @Test
+ public void testPredictOnlyGauss() {
+ GaussianNaiveBayesModel gaussianModel =
+ new GaussianNaiveBayesModel(means, variances, classProbabilities,
labels, null);
+
+ Vector observation = VectorUtils.of(6, 130, 8);
+
+ CompoundNaiveBayesModel model = new CompoundNaiveBayesModel()
+ .wirhPriorProbabilities(classProbabilities)
+ .withLabels(labels)
+ .withGaussianModel(gaussianModel);
+
+ assertEquals(LABEL_2, model.predict(observation), PRECISION);
+ }
+
+ /** Test. */
+ @Test
+ public void testPredictOnlyDiscrete() {
+ DiscreteNaiveBayesModel discreteModel =
+ new DiscreteNaiveBayesModel(probabilities, classProbabilities,
labels, binarizedDataThresholds, null);
+
+ Vector observation = VectorUtils.of(1, 0, 1, 1, 0);
+
+ CompoundNaiveBayesModel model = new CompoundNaiveBayesModel()
+ .wirhPriorProbabilities(classProbabilities)
+ .withLabels(labels)
+ .withDiscreteModel(discreteModel);
+
+ assertEquals(LABEL_2, model.predict(observation), PRECISION);
+ }
+
+ /** Test. */
+ @Test
+ public void testPredictGausAndDiscrete() {
+ DiscreteNaiveBayesModel discreteModel =
+ new DiscreteNaiveBayesModel(probabilities, classProbabilities,
labels, binarizedDataThresholds, null);
+
+ GaussianNaiveBayesModel gaussianModel =
+ new GaussianNaiveBayesModel(means, variances, classProbabilities,
labels, null);
+
+ CompoundNaiveBayesModel model = new CompoundNaiveBayesModel()
+ .wirhPriorProbabilities(classProbabilities)
+ .withLabels(labels)
+ .withGaussianModel(gaussianModel)
+ .withGaussianFeatureIdsToSkip(asList(3, 4, 5, 6, 7))
+ .withDiscreteModel(discreteModel)
+ .withDiscreteFeatureIdsToSkip( asList(0, 1, 2));
+
+ Vector observation = VectorUtils.of(6, 130, 8, 1, 0, 1, 1, 0);
+
+ assertEquals(LABEL_2, model.predict(observation), PRECISION);
+ }
+}
diff --git
a/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/compound/CompoundNaiveBayesTest.java
b/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/compound/CompoundNaiveBayesTest.java
new file mode 100644
index 0000000..c4cef0e
--- /dev/null
+++
b/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/compound/CompoundNaiveBayesTest.java
@@ -0,0 +1,65 @@
+/*
+ * 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.ignite.ml.naivebayes.compound;
+
+import org.apache.ignite.ml.dataset.feature.extractor.Vectorizer;
+import
org.apache.ignite.ml.dataset.feature.extractor.impl.DoubleArrayVectorizer;
+import org.apache.ignite.ml.dataset.impl.local.LocalDatasetBuilder;
+import org.apache.ignite.ml.math.primitives.vector.Vector;
+import org.apache.ignite.ml.math.primitives.vector.VectorUtils;
+import org.apache.ignite.ml.naivebayes.discrete.DiscreteNaiveBayesTrainer;
+import org.apache.ignite.ml.naivebayes.gaussian.GaussianNaiveBayesTrainer;
+import org.junit.Test;
+
+import static java.util.Arrays.asList;
+import static org.apache.ignite.ml.naivebayes.compound.Data.LABEL_1;
+import static org.apache.ignite.ml.naivebayes.compound.Data.LABEL_2;
+import static
org.apache.ignite.ml.naivebayes.compound.Data.binarizedDataThresholds;
+import static org.apache.ignite.ml.naivebayes.compound.Data.classProbabilities;
+import static org.apache.ignite.ml.naivebayes.compound.Data.data;
+import static org.junit.Assert.assertEquals;
+
+/** Integration tests for Compound naive Bayes algorithm with different
datasets. */
+public class CompoundNaiveBayesTest {
+
+ /** Precision in test checks. */
+ private static final double PRECISION = 1e-2;
+
+ /** Test. */
+ @Test
+ public void testLearnsAndPredictCorrently() {
+ CompoundNaiveBayesTrainer trainer = new CompoundNaiveBayesTrainer()
+ .withPriorProbabilities(classProbabilities)
+ .withGaussianNaiveBayesTrainer(new GaussianNaiveBayesTrainer())
+ .withGaussianFeatureIdsToSkip(asList(3, 4, 5, 6, 7))
+ .withDiscreteNaiveBayesTrainer(new DiscreteNaiveBayesTrainer()
+ .setBucketThresholds(binarizedDataThresholds))
+ .withDiscreteFeatureIdsToSkip(asList(0, 1, 2));
+
+ CompoundNaiveBayesModel model = trainer.fit(
+ new LocalDatasetBuilder<>(data, 2),
+ new
DoubleArrayVectorizer<Integer>().labeled(Vectorizer.LabelCoordinate.LAST)
+ );
+
+ Vector observation1 = VectorUtils.of(5.92, 165, 10, 1, 1, 0, 0, 0);
+ assertEquals(LABEL_1, model.predict(observation1), PRECISION);
+
+ Vector observation2 = VectorUtils.of(6, 130, 8, 1, 0, 1, 1, 0);
+ assertEquals(LABEL_2, model.predict(observation2), PRECISION);
+ }
+}
diff --git
a/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/compound/CompoundNaiveBayesTrainerTest.java
b/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/compound/CompoundNaiveBayesTrainerTest.java
new file mode 100644
index 0000000..9b40fbb
--- /dev/null
+++
b/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/compound/CompoundNaiveBayesTrainerTest.java
@@ -0,0 +1,108 @@
+/*
+ * 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.ignite.ml.naivebayes.compound;
+
+import org.apache.ignite.ml.common.TrainerTest;
+import org.apache.ignite.ml.dataset.feature.extractor.Vectorizer;
+import
org.apache.ignite.ml.dataset.feature.extractor.impl.DoubleArrayVectorizer;
+import org.apache.ignite.ml.dataset.impl.local.LocalDatasetBuilder;
+import org.apache.ignite.ml.naivebayes.discrete.DiscreteNaiveBayesModel;
+import org.apache.ignite.ml.naivebayes.discrete.DiscreteNaiveBayesTrainer;
+import org.apache.ignite.ml.naivebayes.gaussian.GaussianNaiveBayesModel;
+import org.apache.ignite.ml.naivebayes.gaussian.GaussianNaiveBayesTrainer;
+import org.junit.Before;
+import org.junit.Test;
+
+import static java.util.Arrays.asList;
+import static
org.apache.ignite.ml.naivebayes.compound.Data.binarizedDataThresholds;
+import static org.apache.ignite.ml.naivebayes.compound.Data.classProbabilities;
+import static org.apache.ignite.ml.naivebayes.compound.Data.data;
+import static org.junit.Assert.assertArrayEquals;
+import static org.junit.Assert.assertEquals;
+
+/** Test for {@link CompoundNaiveBayesTrainer} */
+public class CompoundNaiveBayesTrainerTest extends TrainerTest {
+
+ /** Precision in test checks. */
+ private static final double PRECISION = 1e-2;
+
+ /** Trainer under test. */
+ private CompoundNaiveBayesTrainer trainer;
+
+ /** Initialization {@code CompoundNaiveBayesTrainer}. */
+ @Before
+ public void createTrainer() {
+ trainer = new CompoundNaiveBayesTrainer()
+ .withPriorProbabilities(classProbabilities)
+ .withGaussianNaiveBayesTrainer(new GaussianNaiveBayesTrainer())
+ .withGaussianFeatureIdsToSkip(asList(3, 4, 5, 6, 7))
+ .withDiscreteNaiveBayesTrainer(new DiscreteNaiveBayesTrainer()
+ .setBucketThresholds(binarizedDataThresholds))
+ .withDiscreteFeatureIdsToSkip(asList(0, 1, 2));
+ }
+
+ /** Test. */
+ @Test
+ public void test() {
+ CompoundNaiveBayesModel model = trainer.fit(
+ new LocalDatasetBuilder<>(data, parts),
+ new
DoubleArrayVectorizer<Integer>().labeled(Vectorizer.LabelCoordinate.LAST)
+ );
+
+ assertDiscreteModel(model.getDiscreteModel());
+ assertGaussianModel(model.getGaussianModel());
+ }
+
+ /** Discrete model assertions. */
+ private void assertDiscreteModel(DiscreteNaiveBayesModel model) {
+ double[][][] expectedProbabilites = new double[][][] {
+ {
+ {.25, .75},
+ {.5, .5},
+ {.5, .5},
+ {.5, .5},
+ {.5, .5}
+ },
+ {
+ {.0, 1},
+ {.25, .75},
+ {.75, .25},
+ {.25, .75},
+ {.5, .5}
+ }
+ };
+
+ for (int i = 0; i < expectedProbabilites.length; i++) {
+ for (int j = 0; j < expectedProbabilites[i].length; j++)
+ assertArrayEquals(expectedProbabilites[i][j],
model.getProbabilities()[i][j], PRECISION);
+ }
+ assertArrayEquals(new double[] {.5, .5}, model.getClsProbabilities(),
PRECISION);
+ }
+
+ /** Gaussian model assertions. */
+ private void assertGaussianModel(GaussianNaiveBayesModel model) {
+ double[] priorProbabilities = new double[] {.5, .5};
+
+ assertEquals(priorProbabilities[0], model.getClassProbabilities()[0],
PRECISION);
+ assertEquals(priorProbabilities[1], model.getClassProbabilities()[1],
PRECISION);
+ assertArrayEquals(new double[] {5.855, 176.25, 11.25},
model.getMeans()[0], PRECISION);
+ assertArrayEquals(new double[] {5.4175, 132.5, 7.5},
model.getMeans()[1], PRECISION);
+ double[] expectedVars = {0.026274999999999, 92.1875, 0.6875};
+ assertArrayEquals(expectedVars, model.getVariances()[0], PRECISION);
+ }
+}
diff --git
a/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/compound/Data.java
b/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/compound/Data.java
new file mode 100644
index 0000000..ee34f85
--- /dev/null
+++
b/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/compound/Data.java
@@ -0,0 +1,88 @@
+/*
+ * 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.ignite.ml.naivebayes.compound;
+
+import java.util.HashMap;
+import java.util.Map;
+
+/** Data class which contains test data with precalculated statistics. */
+final class Data {
+
+ /** Private constructor. */
+ private Data() {
+ }
+
+ /** The first label. */
+ static final double LABEL_1 = 1.;
+
+ /** The second label. */
+ static final double LABEL_2 = 2.;
+
+ /** Labels. */
+ static final double[] labels = {LABEL_1, LABEL_2};
+
+ /** */
+ static final Map<Integer, double[]> data = new HashMap<>();
+
+ /** Means for gaussian data part. */
+ static double[][] means;
+
+ /** Variances for gaussian data part. */
+ static double[][] variances;
+
+ /** */
+ static double[] classProbabilities;
+
+ /** Thresholds to binarize discret data. */
+ static double[][] binarizedDataThresholds;
+
+ /** Discrete probabilities. */
+ static double[][][] probabilities;
+
+ static {
+ data.put(0, new double[] {6, 180, 12, 0, 0, 1, 1, 1, LABEL_1});
+ data.put(1, new double[] {5.92, 190, 11, 1, 0, 1, 1, 0, LABEL_1});
+ data.put(2, new double[] {5.58, 170, 12, 1, 1, 0, 0, 1, LABEL_1});
+ data.put(3, new double[] {5.92, 165, 10, 1, 1, 0, 0, 0, LABEL_1});
+
+ data.put(4, new double[] {5, 100, 6, 1, 0, 0, 1, 1, LABEL_2});
+ data.put(5, new double[] {5.5, 150, 8, 1, 1, 0, 0, 1, LABEL_2});
+ data.put(6, new double[] {5.42, 130, 7, 1, 1, 1, 1, 0, LABEL_2});
+ data.put(7, new double[] {5.75, 150, 9, 1, 1, 0, 1, 0, LABEL_2});
+
+ classProbabilities = new double[] {.5, .5};
+
+ means = new double[][] {
+ {5.855, 176.25, 11.25},
+ {5.4175, 132.5, 7.5},
+ };
+
+ variances = new double[][] {
+ {3.5033E-2, 1.2292E2, 9.1667E-1},
+ {9.7225E-2, 5.5833E2, 1.6667},
+ };
+
+ binarizedDataThresholds = new double[][] {{.5}, {.5}, {.5}, {.5},
{.5}};
+
+ probabilities = new double[][][] {
+ {{.25, .75}, {.25, .75}, {.5, .5}, {.5, .5}, {.5, .5}},
+ {{0, 1}, {.25, .75}, {.75, .25}, {.25, .75}, {.5, .5}}
+ };
+ }
+
+}
diff --git
a/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/discrete/DiscreteNaiveBayesModelTest.java
b/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/discrete/DiscreteNaiveBayesModelTest.java
index 41d320d..60d0cdd 100644
---
a/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/discrete/DiscreteNaiveBayesModelTest.java
+++
b/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/discrete/DiscreteNaiveBayesModelTest.java
@@ -24,7 +24,8 @@ import org.junit.Test;
/** Tests for {@code DiscreteNaiveBayesModel} */
public class DiscreteNaiveBayesModelTest {
- /** */
+
+ /** Test. */
@Test
public void testPredictWithTwoClasses() {
double first = 1;
@@ -36,7 +37,8 @@ public class DiscreteNaiveBayesModelTest {
double[] classProbabilities = new double[] {6. / 13, 7. / 13};
double[][] thresholds = new double[][] {{.5}, {.2, .7}, {.5}, {.5,
1.5}, {.5}};
- DiscreteNaiveBayesModel mdl = new
DiscreteNaiveBayesModel(probabilities, classProbabilities, new double[] {first,
second}, thresholds, new DiscreteNaiveBayesSumsHolder());
+ DiscreteNaiveBayesModel mdl = new
DiscreteNaiveBayesModel(probabilities, classProbabilities,
+ new double[] {first, second}, thresholds, new
DiscreteNaiveBayesSumsHolder());
Vector observation = VectorUtils.of(2, 0, 1, 2, 0);
Assert.assertEquals(second, mdl.predict(observation), 0.0001);
diff --git
a/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/discrete/DiscreteNaiveBayesTrainerTest.java
b/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/discrete/DiscreteNaiveBayesTrainerTest.java
index a864fd6..7426f56 100644
---
a/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/discrete/DiscreteNaiveBayesTrainerTest.java
+++
b/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/discrete/DiscreteNaiveBayesTrainerTest.java
@@ -79,7 +79,7 @@ public class DiscreteNaiveBayesTrainerTest extends
TrainerTest {
data.put(11, new double[] {2, 1, .45, 748, LABEL_2});
}
- /** */
+ /** Trainer under test. */
private DiscreteNaiveBayesTrainer trainer;
/** Initialization {@code DiscreteNaiveBayesTrainer}. */
@@ -88,7 +88,7 @@ public class DiscreteNaiveBayesTrainerTest extends
TrainerTest {
trainer = new
DiscreteNaiveBayesTrainer().setBucketThresholds(binarizedDatathresholds);
}
- /** */
+ /** Test. */
@Test
public void testReturnsCorrectLabelProbalities() {
@@ -101,7 +101,7 @@ public class DiscreteNaiveBayesTrainerTest extends
TrainerTest {
Assert.assertArrayEquals(expectedProbabilities,
model.getClsProbabilities(), PRECISION);
}
- /** */
+ /** Test. */
@Test
public void testReturnsEquivalentProbalitiesWhenSetEquiprobableClasses_() {
DiscreteNaiveBayesTrainer trainer = new DiscreteNaiveBayesTrainer()
@@ -116,7 +116,7 @@ public class DiscreteNaiveBayesTrainerTest extends
TrainerTest {
Assert.assertArrayEquals(new double[] {.5, .5},
model.getClsProbabilities(), PRECISION);
}
- /** */
+ /** Test. */
@Test
public void testReturnsPresetProbalitiesWhenSetPriorProbabilities() {
double[] priorProbabilities = new double[] {.35, .65};
@@ -132,7 +132,7 @@ public class DiscreteNaiveBayesTrainerTest extends
TrainerTest {
Assert.assertArrayEquals(priorProbabilities,
model.getClsProbabilities(), PRECISION);
}
- /** */
+ /** Test. */
@Test
public void testReturnsCorrectPriorProbabilities() {
double[][][] expectedPriorProbabilites = new double[][][] {
@@ -151,7 +151,7 @@ public class DiscreteNaiveBayesTrainerTest extends
TrainerTest {
}
}
- /** */
+ /** Test. */
@Test
public void testReturnsCorrectPriorProbabilitiesWithDefferentThresholds() {
double[][][] expectedPriorProbabilites = new double[][][] {
diff --git
a/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/gaussian/GaussianNaiveBayesModelTest.java
b/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/gaussian/GaussianNaiveBayesModelTest.java
index d35ea3d..043ef38 100644
---
a/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/gaussian/GaussianNaiveBayesModelTest.java
+++
b/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/gaussian/GaussianNaiveBayesModelTest.java
@@ -26,7 +26,8 @@ import org.junit.Test;
* Tests for {@link GaussianNaiveBayesModel}.
*/
public class GaussianNaiveBayesModelTest {
- /** */
+
+ /** Test. */
@Test
public void testPredictWithTwoClasses() {
double first = 1;
diff --git
a/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/gaussian/GaussianNaiveBayesTrainerTest.java
b/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/gaussian/GaussianNaiveBayesTrainerTest.java
index 80927e9..d00cfc6 100644
---
a/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/gaussian/GaussianNaiveBayesTrainerTest.java
+++
b/modules/ml/src/test/java/org/apache/ignite/ml/naivebayes/gaussian/GaussianNaiveBayesTrainerTest.java
@@ -36,19 +36,19 @@ public class GaussianNaiveBayesTrainerTest extends
TrainerTest {
/** Precision in test checks. */
private static final double PRECISION = 1e-2;
- /** */
+ /** Label. */
private static final double LABEL_1 = 1.;
- /** */
+ /** Label. */
private static final double LABEL_2 = 2.;
/** Data. */
private static final Map<Integer, double[]> data = new HashMap<>();
- /** */
+ /** {@code LABEL_1} data. */
private static final Map<Integer, double[]> singleLabeldata1 = new
HashMap<>();
- /** */
+ /** {@code LABEL_2} data. */
private static final Map<Integer, double[]> singleLabeldata2 = new
HashMap<>();
static {
@@ -75,7 +75,7 @@ public class GaussianNaiveBayesTrainerTest extends
TrainerTest {
trainer = new GaussianNaiveBayesTrainer();
}
- /** */
+ /** Test. */
@Test
public void testWithLinearlySeparableData() {
Map<Integer, double[]> cacheMock = new HashMap<>();
@@ -92,7 +92,7 @@ public class GaussianNaiveBayesTrainerTest extends
TrainerTest {
TestUtils.assertEquals(1, mdl.predict(VectorUtils.of(10, 100)),
PRECISION);
}
- /** */
+ /** Test. */
@Test
public void testReturnsCorrectLabelProbalities() {
@@ -105,7 +105,7 @@ public class GaussianNaiveBayesTrainerTest extends
TrainerTest {
Assert.assertEquals(2. / data.size(),
model.getClassProbabilities()[1], PRECISION);
}
- /** */
+ /** Test. */
@Test
public void testReturnsEquivalentProbalitiesWhenSetEquiprobableClasses_() {
GaussianNaiveBayesTrainer trainer = new GaussianNaiveBayesTrainer()
@@ -120,7 +120,7 @@ public class GaussianNaiveBayesTrainerTest extends
TrainerTest {
Assert.assertEquals(.5, model.getClassProbabilities()[1], PRECISION);
}
- /** */
+ /** Test. */
@Test
public void testReturnsPresetProbalitiesWhenSetPriorProbabilities() {
double[] priorProbabilities = new double[] {.35, .65};
@@ -136,7 +136,7 @@ public class GaussianNaiveBayesTrainerTest extends
TrainerTest {
Assert.assertEquals(priorProbabilities[1],
model.getClassProbabilities()[1], PRECISION);
}
- /** */
+ /** Test. */
@Test
public void testReturnsCorrectMeans() {
@@ -148,7 +148,7 @@ public class GaussianNaiveBayesTrainerTest extends
TrainerTest {
Assert.assertArrayEquals(new double[] {2.0, 2. / 3.},
model.getMeans()[0], PRECISION);
}
- /** */
+ /** Test. */
@Test
public void testReturnsCorrectVariances() {
@@ -161,7 +161,7 @@ public class GaussianNaiveBayesTrainerTest extends
TrainerTest {
Assert.assertArrayEquals(expectedVars, model.getVariances()[0],
PRECISION);
}
- /** */
+ /** Test. */
@Test
public void testUpdatigModel() {
Vectorizer<Integer, double[], Integer, Double> vectorizer = new
DoubleArrayVectorizer<Integer>().labeled(Vectorizer.LabelCoordinate.LAST);
