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
|
/*
* 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.feature
import org.apache.hadoop.fs.Path
import org.apache.spark.annotation.Since
import org.apache.spark.ml._
import org.apache.spark.ml.attribute.{AttributeGroup, _}
import org.apache.spark.ml.linalg.{Vector, VectorUDT}
import org.apache.spark.ml.param._
import org.apache.spark.ml.param.shared._
import org.apache.spark.ml.util._
import org.apache.spark.mllib.feature
import org.apache.spark.mllib.linalg.{Vectors => OldVectors}
import org.apache.spark.mllib.regression.{LabeledPoint => OldLabeledPoint}
import org.apache.spark.rdd.RDD
import org.apache.spark.sql._
import org.apache.spark.sql.functions._
import org.apache.spark.sql.types.{DoubleType, StructField, StructType}
/**
* Params for [[ChiSqSelector]] and [[ChiSqSelectorModel]].
*/
private[feature] trait ChiSqSelectorParams extends Params
with HasFeaturesCol with HasOutputCol with HasLabelCol {
/**
* Number of features that selector will select (ordered by statistic value descending). If the
* number of features is less than numTopFeatures, then this will select all features.
* The default value of numTopFeatures is 50.
* @group param
*/
final val numTopFeatures = new IntParam(this, "numTopFeatures",
"Number of features that selector will select, ordered by statistics value descending. If the" +
" number of features is < numTopFeatures, then this will select all features.",
ParamValidators.gtEq(1))
setDefault(numTopFeatures -> 50)
/** @group getParam */
def getNumTopFeatures: Int = $(numTopFeatures)
}
/**
* Chi-Squared feature selection, which selects categorical features to use for predicting a
* categorical label.
*/
@Since("1.6.0")
final class ChiSqSelector @Since("1.6.0") (@Since("1.6.0") override val uid: String)
extends Estimator[ChiSqSelectorModel] with ChiSqSelectorParams with DefaultParamsWritable {
@Since("1.6.0")
def this() = this(Identifiable.randomUID("chiSqSelector"))
/** @group setParam */
@Since("1.6.0")
def setNumTopFeatures(value: Int): this.type = set(numTopFeatures, value)
/** @group setParam */
@Since("1.6.0")
def setFeaturesCol(value: String): this.type = set(featuresCol, value)
/** @group setParam */
@Since("1.6.0")
def setOutputCol(value: String): this.type = set(outputCol, value)
/** @group setParam */
@Since("1.6.0")
def setLabelCol(value: String): this.type = set(labelCol, value)
@Since("2.0.0")
override def fit(dataset: Dataset[_]): ChiSqSelectorModel = {
transformSchema(dataset.schema, logging = true)
val input: RDD[OldLabeledPoint] =
dataset.select(col($(labelCol)).cast(DoubleType), col($(featuresCol))).rdd.map {
case Row(label: Double, features: Vector) =>
OldLabeledPoint(label, OldVectors.fromML(features))
}
val chiSqSelector = new feature.ChiSqSelector($(numTopFeatures)).fit(input)
copyValues(new ChiSqSelectorModel(uid, chiSqSelector).setParent(this))
}
@Since("1.6.0")
override def transformSchema(schema: StructType): StructType = {
SchemaUtils.checkColumnType(schema, $(featuresCol), new VectorUDT)
SchemaUtils.checkNumericType(schema, $(labelCol))
SchemaUtils.appendColumn(schema, $(outputCol), new VectorUDT)
}
@Since("1.6.0")
override def copy(extra: ParamMap): ChiSqSelector = defaultCopy(extra)
}
@Since("1.6.0")
object ChiSqSelector extends DefaultParamsReadable[ChiSqSelector] {
@Since("1.6.0")
override def load(path: String): ChiSqSelector = super.load(path)
}
/**
* Model fitted by [[ChiSqSelector]].
*/
@Since("1.6.0")
final class ChiSqSelectorModel private[ml] (
@Since("1.6.0") override val uid: String,
private val chiSqSelector: feature.ChiSqSelectorModel)
extends Model[ChiSqSelectorModel] with ChiSqSelectorParams with MLWritable {
import ChiSqSelectorModel._
/** list of indices to select (filter). Must be ordered asc */
@Since("1.6.0")
val selectedFeatures: Array[Int] = chiSqSelector.selectedFeatures
/** @group setParam */
@Since("1.6.0")
def setFeaturesCol(value: String): this.type = set(featuresCol, value)
/** @group setParam */
@Since("1.6.0")
def setOutputCol(value: String): this.type = set(outputCol, value)
/**
* @group setParam
*/
@Since("1.6.0")
@deprecated("labelCol is not used by ChiSqSelectorModel.", "2.0.0")
def setLabelCol(value: String): this.type = set(labelCol, value)
@Since("2.0.0")
override def transform(dataset: Dataset[_]): DataFrame = {
val transformedSchema = transformSchema(dataset.schema, logging = true)
val newField = transformedSchema.last
// TODO: Make the transformer natively in ml framework to avoid extra conversion.
val transformer: Vector => Vector = v => chiSqSelector.transform(OldVectors.fromML(v)).asML
val selector = udf(transformer)
dataset.withColumn($(outputCol), selector(col($(featuresCol))), newField.metadata)
}
@Since("1.6.0")
override def transformSchema(schema: StructType): StructType = {
SchemaUtils.checkColumnType(schema, $(featuresCol), new VectorUDT)
val newField = prepOutputField(schema)
val outputFields = schema.fields :+ newField
StructType(outputFields)
}
/**
* Prepare the output column field, including per-feature metadata.
*/
private def prepOutputField(schema: StructType): StructField = {
val selector = chiSqSelector.selectedFeatures.toSet
val origAttrGroup = AttributeGroup.fromStructField(schema($(featuresCol)))
val featureAttributes: Array[Attribute] = if (origAttrGroup.attributes.nonEmpty) {
origAttrGroup.attributes.get.zipWithIndex.filter(x => selector.contains(x._2)).map(_._1)
} else {
Array.fill[Attribute](selector.size)(NominalAttribute.defaultAttr)
}
val newAttributeGroup = new AttributeGroup($(outputCol), featureAttributes)
newAttributeGroup.toStructField()
}
@Since("1.6.0")
override def copy(extra: ParamMap): ChiSqSelectorModel = {
val copied = new ChiSqSelectorModel(uid, chiSqSelector)
copyValues(copied, extra).setParent(parent)
}
@Since("1.6.0")
override def write: MLWriter = new ChiSqSelectorModelWriter(this)
}
@Since("1.6.0")
object ChiSqSelectorModel extends MLReadable[ChiSqSelectorModel] {
private[ChiSqSelectorModel]
class ChiSqSelectorModelWriter(instance: ChiSqSelectorModel) extends MLWriter {
private case class Data(selectedFeatures: Seq[Int])
override protected def saveImpl(path: String): Unit = {
DefaultParamsWriter.saveMetadata(instance, path, sc)
val data = Data(instance.selectedFeatures.toSeq)
val dataPath = new Path(path, "data").toString
sparkSession.createDataFrame(Seq(data)).repartition(1).write.parquet(dataPath)
}
}
private class ChiSqSelectorModelReader extends MLReader[ChiSqSelectorModel] {
private val className = classOf[ChiSqSelectorModel].getName
override def load(path: String): ChiSqSelectorModel = {
val metadata = DefaultParamsReader.loadMetadata(path, sc, className)
val dataPath = new Path(path, "data").toString
val data = sparkSession.read.parquet(dataPath).select("selectedFeatures").head()
val selectedFeatures = data.getAs[Seq[Int]](0).toArray
val oldModel = new feature.ChiSqSelectorModel(selectedFeatures)
val model = new ChiSqSelectorModel(metadata.uid, oldModel)
DefaultParamsReader.getAndSetParams(model, metadata)
model
}
}
@Since("1.6.0")
override def read: MLReader[ChiSqSelectorModel] = new ChiSqSelectorModelReader
@Since("1.6.0")
override def load(path: String): ChiSqSelectorModel = super.load(path)
}
|
