1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
|
/*
* 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.spark.ml.classification
import org.apache.spark.SparkFunSuite
import org.apache.spark.ml.impl.TreeTests
import org.apache.spark.mllib.regression.LabeledPoint
import org.apache.spark.mllib.tree.{EnsembleTestHelper, GradientBoostedTrees => OldGBT}
import org.apache.spark.mllib.tree.configuration.{Algo => OldAlgo}
import org.apache.spark.mllib.util.MLlibTestSparkContext
import org.apache.spark.rdd.RDD
import org.apache.spark.sql.DataFrame
/**
* Test suite for [[GBTClassifier]].
*/
class GBTClassifierSuite extends SparkFunSuite with MLlibTestSparkContext {
import GBTClassifierSuite.compareAPIs
// Combinations for estimators, learning rates and subsamplingRate
private val testCombinations =
Array((10, 1.0, 1.0), (10, 0.1, 1.0), (10, 0.5, 0.75), (10, 0.1, 0.75))
private var data: RDD[LabeledPoint] = _
private var trainData: RDD[LabeledPoint] = _
private var validationData: RDD[LabeledPoint] = _
override def beforeAll() {
super.beforeAll()
data = sc.parallelize(EnsembleTestHelper.generateOrderedLabeledPoints(numFeatures = 10, 100), 2)
trainData =
sc.parallelize(EnsembleTestHelper.generateOrderedLabeledPoints(numFeatures = 20, 120), 2)
validationData =
sc.parallelize(EnsembleTestHelper.generateOrderedLabeledPoints(numFeatures = 20, 80), 2)
}
test("Binary classification with continuous features: Log Loss") {
val categoricalFeatures = Map.empty[Int, Int]
testCombinations.foreach {
case (maxIter, learningRate, subsamplingRate) =>
val gbt = new GBTClassifier()
.setMaxDepth(2)
.setSubsamplingRate(subsamplingRate)
.setLossType("logistic")
.setMaxIter(maxIter)
.setStepSize(learningRate)
compareAPIs(data, None, gbt, categoricalFeatures)
}
}
// TODO: Reinstate test once runWithValidation is implemented SPARK-7132
/*
test("runWithValidation stops early and performs better on a validation dataset") {
val categoricalFeatures = Map.empty[Int, Int]
// Set maxIter large enough so that it stops early.
val maxIter = 20
GBTClassifier.supportedLossTypes.foreach { loss =>
val gbt = new GBTClassifier()
.setMaxIter(maxIter)
.setMaxDepth(2)
.setLossType(loss)
.setValidationTol(0.0)
compareAPIs(trainData, None, gbt, categoricalFeatures)
compareAPIs(trainData, Some(validationData), gbt, categoricalFeatures)
}
}
*/
/////////////////////////////////////////////////////////////////////////////
// Tests of model save/load
/////////////////////////////////////////////////////////////////////////////
// TODO: Reinstate test once save/load are implemented SPARK-6725
/*
test("model save/load") {
val tempDir = Utils.createTempDir()
val path = tempDir.toURI.toString
val trees = Range(0, 3).map(_ => OldDecisionTreeSuite.createModel(OldAlgo.Regression)).toArray
val treeWeights = Array(0.1, 0.3, 1.1)
val oldModel = new OldGBTModel(OldAlgo.Classification, trees, treeWeights)
val newModel = GBTClassificationModel.fromOld(oldModel)
// Save model, load it back, and compare.
try {
newModel.save(sc, path)
val sameNewModel = GBTClassificationModel.load(sc, path)
TreeTests.checkEqual(newModel, sameNewModel)
} finally {
Utils.deleteRecursively(tempDir)
}
}
*/
}
private object GBTClassifierSuite {
/**
* Train 2 models on the given dataset, one using the old API and one using the new API.
* Convert the old model to the new format, compare them, and fail if they are not exactly equal.
*/
def compareAPIs(
data: RDD[LabeledPoint],
validationData: Option[RDD[LabeledPoint]],
gbt: GBTClassifier,
categoricalFeatures: Map[Int, Int]): Unit = {
val oldBoostingStrategy =
gbt.getOldBoostingStrategy(categoricalFeatures, OldAlgo.Classification)
val oldGBT = new OldGBT(oldBoostingStrategy)
val oldModel = oldGBT.run(data)
val newData: DataFrame = TreeTests.setMetadata(data, categoricalFeatures, numClasses = 2)
val newModel = gbt.fit(newData)
// Use parent, fittingParamMap from newTree since these are not checked anyways.
val oldModelAsNew = GBTClassificationModel.fromOld(
oldModel, newModel.parent.asInstanceOf[GBTClassifier], categoricalFeatures)
TreeTests.checkEqual(oldModelAsNew, newModel)
}
}
|
