/
NaiveBayesSuite.scala
365 lines (316 loc) · 13.7 KB
/
NaiveBayesSuite.scala
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
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
/*
* 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 scala.util.Random
import breeze.linalg.{DenseVector => BDV, Vector => BV}
import breeze.stats.distributions.{Multinomial => BrzMultinomial, RandBasis => BrzRandBasis}
import org.apache.spark.{SparkException, SparkFunSuite}
import org.apache.spark.ml.classification.NaiveBayes.{Bernoulli, Multinomial}
import org.apache.spark.ml.classification.NaiveBayesSuite._
import org.apache.spark.ml.feature.LabeledPoint
import org.apache.spark.ml.linalg._
import org.apache.spark.ml.param.ParamsSuite
import org.apache.spark.ml.util.{DefaultReadWriteTest, MLTest, MLTestingUtils}
import org.apache.spark.ml.util.TestingUtils._
import org.apache.spark.sql.{DataFrame, Dataset, Row}
class NaiveBayesSuite extends MLTest with DefaultReadWriteTest {
import testImplicits._
@transient var dataset: Dataset[_] = _
@transient var bernoulliDataset: Dataset[_] = _
private val seed = 42
override def beforeAll(): Unit = {
super.beforeAll()
val pi = Array(0.3, 0.3, 0.4).map(math.log)
val theta = Array(
Array(0.30, 0.30, 0.30, 0.30), // label 0
Array(0.30, 0.30, 0.30, 0.30), // label 1
Array(0.40, 0.40, 0.40, 0.40) // label 2
).map(_.map(math.log))
dataset = generateNaiveBayesInput(pi, theta, 100, seed).toDF()
bernoulliDataset = generateNaiveBayesInput(pi, theta, 100, seed, "bernoulli").toDF()
}
def validatePrediction(predictionAndLabels: Seq[Row]): Unit = {
val numOfErrorPredictions = predictionAndLabels.filter {
case Row(prediction: Double, label: Double) =>
prediction != label
}.length
// At least 80% of the predictions should be on.
assert(numOfErrorPredictions < predictionAndLabels.length / 5)
}
def validateModelFit(
piData: Vector,
thetaData: Matrix,
model: NaiveBayesModel): Unit = {
assert(Vectors.dense(model.pi.toArray.map(math.exp)) ~==
Vectors.dense(piData.toArray.map(math.exp)) absTol 0.05, "pi mismatch")
assert(model.theta.map(math.exp) ~== thetaData.map(math.exp) absTol 0.05, "theta mismatch")
}
def expectedMultinomialProbabilities(model: NaiveBayesModel, feature: Vector): Vector = {
val logClassProbs: BV[Double] = model.pi.asBreeze + model.theta.multiply(feature).asBreeze
val classProbs = logClassProbs.toArray.map(math.exp)
val classProbsSum = classProbs.sum
Vectors.dense(classProbs.map(_ / classProbsSum))
}
def expectedBernoulliProbabilities(model: NaiveBayesModel, feature: Vector): Vector = {
val negThetaMatrix = model.theta.map(v => math.log(1.0 - math.exp(v)))
val negFeature = Vectors.dense(feature.toArray.map(v => 1.0 - v))
val piTheta: BV[Double] = model.pi.asBreeze + model.theta.multiply(feature).asBreeze
val logClassProbs: BV[Double] = piTheta + negThetaMatrix.multiply(negFeature).asBreeze
val classProbs = logClassProbs.toArray.map(math.exp)
val classProbsSum = classProbs.sum
Vectors.dense(classProbs.map(_ / classProbsSum))
}
def validateProbabilities(
featureAndProbabilities: Seq[Row],
model: NaiveBayesModel,
modelType: String): Unit = {
featureAndProbabilities.foreach {
case Row(features: Vector, probability: Vector) =>
assert(probability.toArray.sum ~== 1.0 relTol 1.0e-10)
val expected = modelType match {
case Multinomial =>
expectedMultinomialProbabilities(model, features)
case Bernoulli =>
expectedBernoulliProbabilities(model, features)
case _ =>
throw new UnknownError(s"Invalid modelType: $modelType.")
}
assert(probability ~== expected relTol 1.0e-10)
}
}
test("model types") {
assert(Multinomial === "multinomial")
assert(Bernoulli === "bernoulli")
}
test("params") {
ParamsSuite.checkParams(new NaiveBayes)
val model = new NaiveBayesModel("nb", pi = Vectors.dense(Array(0.2, 0.8)),
theta = new DenseMatrix(2, 3, Array(0.1, 0.2, 0.3, 0.4, 0.6, 0.4)))
ParamsSuite.checkParams(model)
}
test("naive bayes: default params") {
val nb = new NaiveBayes
assert(nb.getLabelCol === "label")
assert(nb.getFeaturesCol === "features")
assert(nb.getPredictionCol === "prediction")
assert(nb.getSmoothing === 1.0)
assert(nb.getModelType === "multinomial")
}
test("Naive Bayes Multinomial") {
val nPoints = 1000
val piArray = Array(0.5, 0.1, 0.4).map(math.log)
val thetaArray = Array(
Array(0.70, 0.10, 0.10, 0.10), // label 0
Array(0.10, 0.70, 0.10, 0.10), // label 1
Array(0.10, 0.10, 0.70, 0.10) // label 2
).map(_.map(math.log))
val pi = Vectors.dense(piArray)
val theta = new DenseMatrix(3, 4, thetaArray.flatten, true)
val testDataset =
generateNaiveBayesInput(piArray, thetaArray, nPoints, seed, "multinomial").toDF()
val nb = new NaiveBayes().setSmoothing(1.0).setModelType("multinomial")
val model = nb.fit(testDataset)
validateModelFit(pi, theta, model)
assert(model.hasParent)
MLTestingUtils.checkCopyAndUids(nb, model)
val validationDataset =
generateNaiveBayesInput(piArray, thetaArray, nPoints, 17, "multinomial").toDF()
testTransformerByGlobalCheckFunc[(Double, Vector)](validationDataset, model,
"prediction", "label") { predictionAndLabels: Seq[Row] =>
validatePrediction(predictionAndLabels)
}
testTransformerByGlobalCheckFunc[(Double, Vector)](validationDataset, model,
"features", "probability") { featureAndProbabilities: Seq[Row] =>
validateProbabilities(featureAndProbabilities, model, "multinomial")
}
ProbabilisticClassifierSuite.testPredictMethods[
Vector, NaiveBayesModel](this, model, testDataset)
}
test("Naive Bayes with weighted samples") {
val numClasses = 3
def modelEquals(m1: NaiveBayesModel, m2: NaiveBayesModel): Unit = {
assert(m1.pi ~== m2.pi relTol 0.01)
assert(m1.theta ~== m2.theta relTol 0.01)
}
val testParams = Seq[(String, Dataset[_])](
("bernoulli", bernoulliDataset),
("multinomial", dataset)
)
testParams.foreach { case (family, dataset) =>
// NaiveBayes is sensitive to constant scaling of the weights unless smoothing is set to 0
val estimatorNoSmoothing = new NaiveBayes().setSmoothing(0.0).setModelType(family)
val estimatorWithSmoothing = new NaiveBayes().setModelType(family)
MLTestingUtils.testArbitrarilyScaledWeights[NaiveBayesModel, NaiveBayes](
dataset.as[LabeledPoint], estimatorNoSmoothing, modelEquals)
MLTestingUtils.testOutliersWithSmallWeights[NaiveBayesModel, NaiveBayes](
dataset.as[LabeledPoint], estimatorWithSmoothing, numClasses, modelEquals, outlierRatio = 3)
MLTestingUtils.testOversamplingVsWeighting[NaiveBayesModel, NaiveBayes](
dataset.as[LabeledPoint], estimatorWithSmoothing, modelEquals, seed)
}
}
test("Naive Bayes Bernoulli") {
val nPoints = 10000
val piArray = Array(0.5, 0.3, 0.2).map(math.log)
val thetaArray = Array(
Array(0.50, 0.02, 0.02, 0.02, 0.02, 0.02, 0.02, 0.02, 0.02, 0.02, 0.02, 0.40), // label 0
Array(0.02, 0.70, 0.10, 0.02, 0.02, 0.02, 0.02, 0.02, 0.02, 0.02, 0.02, 0.02), // label 1
Array(0.02, 0.02, 0.60, 0.02, 0.02, 0.02, 0.02, 0.02, 0.02, 0.02, 0.02, 0.30) // label 2
).map(_.map(math.log))
val pi = Vectors.dense(piArray)
val theta = new DenseMatrix(3, 12, thetaArray.flatten, true)
val testDataset =
generateNaiveBayesInput(piArray, thetaArray, nPoints, 45, "bernoulli").toDF()
val nb = new NaiveBayes().setSmoothing(1.0).setModelType("bernoulli")
val model = nb.fit(testDataset)
validateModelFit(pi, theta, model)
assert(model.hasParent)
val validationDataset =
generateNaiveBayesInput(piArray, thetaArray, nPoints, 20, "bernoulli").toDF()
testTransformerByGlobalCheckFunc[(Double, Vector)](validationDataset, model,
"prediction", "label") { predictionAndLabels: Seq[Row] =>
validatePrediction(predictionAndLabels)
}
testTransformerByGlobalCheckFunc[(Double, Vector)](validationDataset, model,
"features", "probability") { featureAndProbabilities: Seq[Row] =>
validateProbabilities(featureAndProbabilities, model, "bernoulli")
}
ProbabilisticClassifierSuite.testPredictMethods[
Vector, NaiveBayesModel](this, model, testDataset)
}
test("detect negative values") {
val dense = spark.createDataFrame(Seq(
LabeledPoint(1.0, Vectors.dense(1.0)),
LabeledPoint(0.0, Vectors.dense(-1.0)),
LabeledPoint(1.0, Vectors.dense(1.0)),
LabeledPoint(1.0, Vectors.dense(0.0))))
intercept[SparkException] {
new NaiveBayes().fit(dense)
}
val sparse = spark.createDataFrame(Seq(
LabeledPoint(1.0, Vectors.sparse(1, Array(0), Array(1.0))),
LabeledPoint(0.0, Vectors.sparse(1, Array(0), Array(-1.0))),
LabeledPoint(1.0, Vectors.sparse(1, Array(0), Array(1.0))),
LabeledPoint(1.0, Vectors.sparse(1, Array.empty, Array.empty))))
intercept[SparkException] {
new NaiveBayes().fit(sparse)
}
val nan = spark.createDataFrame(Seq(
LabeledPoint(1.0, Vectors.sparse(1, Array(0), Array(1.0))),
LabeledPoint(0.0, Vectors.sparse(1, Array(0), Array(Double.NaN))),
LabeledPoint(1.0, Vectors.sparse(1, Array(0), Array(1.0))),
LabeledPoint(1.0, Vectors.sparse(1, Array.empty, Array.empty))))
intercept[SparkException] {
new NaiveBayes().fit(nan)
}
}
test("detect non zero or one values in Bernoulli") {
val badTrain = spark.createDataFrame(Seq(
LabeledPoint(1.0, Vectors.dense(1.0)),
LabeledPoint(0.0, Vectors.dense(2.0)),
LabeledPoint(1.0, Vectors.dense(1.0)),
LabeledPoint(1.0, Vectors.dense(0.0))))
intercept[SparkException] {
new NaiveBayes().setModelType(Bernoulli).setSmoothing(1.0).fit(badTrain)
}
val okTrain = spark.createDataFrame(Seq(
LabeledPoint(1.0, Vectors.dense(1.0)),
LabeledPoint(0.0, Vectors.dense(0.0)),
LabeledPoint(1.0, Vectors.dense(1.0)),
LabeledPoint(1.0, Vectors.dense(1.0)),
LabeledPoint(0.0, Vectors.dense(0.0)),
LabeledPoint(1.0, Vectors.dense(1.0)),
LabeledPoint(1.0, Vectors.dense(1.0))))
val model = new NaiveBayes().setModelType(Bernoulli).setSmoothing(1.0).fit(okTrain)
val badPredict = spark.createDataFrame(Seq(
LabeledPoint(1.0, Vectors.dense(1.0)),
LabeledPoint(1.0, Vectors.dense(2.0)),
LabeledPoint(1.0, Vectors.dense(1.0)),
LabeledPoint(1.0, Vectors.dense(0.0))))
intercept[SparkException] {
model.transform(badPredict).collect()
}
}
test("read/write") {
def checkModelData(model: NaiveBayesModel, model2: NaiveBayesModel): Unit = {
assert(model.pi === model2.pi)
assert(model.theta === model2.theta)
}
val nb = new NaiveBayes()
testEstimatorAndModelReadWrite(nb, dataset, NaiveBayesSuite.allParamSettings,
NaiveBayesSuite.allParamSettings, checkModelData)
}
test("should support all NumericType labels and weights, and not support other types") {
val nb = new NaiveBayes()
MLTestingUtils.checkNumericTypes[NaiveBayesModel, NaiveBayes](
nb, spark) { (expected, actual) =>
assert(expected.pi === actual.pi)
assert(expected.theta === actual.theta)
}
}
}
object NaiveBayesSuite {
/**
* Mapping from all Params to valid settings which differ from the defaults.
* This is useful for tests which need to exercise all Params, such as save/load.
* This excludes input columns to simplify some tests.
*/
val allParamSettings: Map[String, Any] = Map(
"predictionCol" -> "myPrediction",
"smoothing" -> 0.1
)
private def calcLabel(p: Double, pi: Array[Double]): Int = {
var sum = 0.0
for (j <- 0 until pi.length) {
sum += pi(j)
if (p < sum) return j
}
-1
}
// Generate input of the form Y = (theta * x).argmax()
def generateNaiveBayesInput(
pi: Array[Double], // 1XC
theta: Array[Array[Double]], // CXD
nPoints: Int,
seed: Int,
modelType: String = Multinomial,
sample: Int = 10): Seq[LabeledPoint] = {
val D = theta(0).length
val rnd = new Random(seed)
val _pi = pi.map(math.exp)
val _theta = theta.map(row => row.map(math.exp))
implicit val rngForBrzMultinomial = BrzRandBasis.withSeed(seed)
for (i <- 0 until nPoints) yield {
val y = calcLabel(rnd.nextDouble(), _pi)
val xi = modelType match {
case Bernoulli => Array.tabulate[Double] (D) { j =>
if (rnd.nextDouble () < _theta(y)(j) ) 1 else 0
}
case Multinomial =>
val mult = BrzMultinomial(BDV(_theta(y)))
val emptyMap = (0 until D).map(x => (x, 0.0)).toMap
val counts = emptyMap ++ mult.sample(sample).groupBy(x => x).map {
case (index, reps) => (index, reps.size.toDouble)
}
counts.toArray.sortBy(_._1).map(_._2)
case _ =>
// This should never happen.
throw new UnknownError(s"Invalid modelType: $modelType.")
}
LabeledPoint(y, Vectors.dense(xi))
}
}
}