Github user MLnick commented on a diff in the pull request:
https://github.com/apache/spark/pull/11601#discussion_r104404523
--- Diff: mllib/src/main/scala/org/apache/spark/ml/feature/Imputer.scala ---
@@ -0,0 +1,260 @@
+/*
+ * 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.SparkException
+import org.apache.spark.annotation.{Experimental, Since}
+import org.apache.spark.ml.{Estimator, Model}
+import org.apache.spark.ml.param._
+import org.apache.spark.ml.param.shared.HasInputCols
+import org.apache.spark.ml.util._
+import org.apache.spark.sql.{DataFrame, Dataset, Row}
+import org.apache.spark.sql.functions._
+import org.apache.spark.sql.types._
+
+/**
+ * Params for [[Imputer]] and [[ImputerModel]].
+ */
+private[feature] trait ImputerParams extends Params with HasInputCols {
+
+ /**
+ * The imputation strategy.
+ * If "mean", then replace missing values using the mean value of the
feature.
+ * If "median", then replace missing values using the approximate median
value of the
+ * feature (relative error less than 0.001).
+ * Default: mean
+ *
+ * @group param
+ */
+ final val strategy: Param[String] = new Param(this, "strategy",
s"strategy for imputation. " +
+ s"If ${Imputer.mean}, then replace missing values using the mean value
of the feature. " +
+ s"If ${Imputer.median}, then replace missing values using the median
value of the feature.",
+ ParamValidators.inArray[String](Array(Imputer.mean, Imputer.median)))
+
+ /** @group getParam */
+ def getStrategy: String = $(strategy)
+
+ /**
+ * The placeholder for the missing values. All occurrences of
missingValue will be imputed.
+ * Note that null values are always treated as missing.
+ * Default: Double.NaN
+ *
+ * @group param
+ */
+ final val missingValue: DoubleParam = new DoubleParam(this,
"missingValue",
+ "The placeholder for the missing values. All occurrences of
missingValue will be imputed")
+
+ /** @group getParam */
+ def getMissingValue: Double = $(missingValue)
+
+ /**
+ * Param for output column names.
+ * @group param
+ */
+ final val outputCols: StringArrayParam = new StringArrayParam(this,
"outputCols",
+ "output column names")
+
+ /** @group getParam */
+ final def getOutputCols: Array[String] = $(outputCols)
+
+ /** Validates and transforms the input schema. */
+ protected def validateAndTransformSchema(schema: StructType): StructType
= {
+ require($(inputCols).length == $(inputCols).distinct.length,
s"inputCols duplicates:" +
+ s" (${$(inputCols).mkString(", ")})")
+ require($(outputCols).length == $(outputCols).distinct.length,
s"outputCols duplicates:" +
+ s" (${$(outputCols).mkString(", ")})")
+ require($(inputCols).length == $(outputCols).length,
s"inputCols(${$(inputCols).length})" +
+ s" and outputCols(${$(outputCols).length}) should have the same
length")
+ val outputFields = $(inputCols).zip($(outputCols)).map { case
(inputCol, outputCol) =>
+ val inputField = schema(inputCol)
+ SchemaUtils.checkColumnTypes(schema, inputCol, Seq(DoubleType,
FloatType))
+ StructField(outputCol, inputField.dataType, inputField.nullable)
+ }
+ StructType(schema ++ outputFields)
+ }
+}
+
+/**
+ * :: Experimental ::
+ * Imputation estimator for completing missing values, either using the
mean or the median
+ * of the column in which the missing values are located. The input column
should be of
+ * DoubleType or FloatType. Currently Imputer does not support categorical
features yet
+ * (SPARK-15041) and possibly creates incorrect values for a categorical
feature.
+ *
+ * Note that the mean/median value is computed after filtering out missing
values.
+ * All Null values in the input column are treated as missing, and so are
also imputed.
+ */
+@Experimental
+class Imputer @Since("2.2.0")(override val uid: String)
+ extends Estimator[ImputerModel] with ImputerParams with
DefaultParamsWritable {
+
+ @Since("2.2.0")
+ def this() = this(Identifiable.randomUID("imputer"))
+
+ /** @group setParam */
+ @Since("2.2.0")
+ def setInputCols(value: Array[String]): this.type = set(inputCols, value)
+
+ /** @group setParam */
+ @Since("2.2.0")
+ def setOutputCols(value: Array[String]): this.type = set(outputCols,
value)
+
+ /**
+ * Imputation strategy. Available options are ["mean", "median"].
+ * @group setParam
+ */
+ @Since("2.2.0")
+ def setStrategy(value: String): this.type = set(strategy, value)
+
+ /** @group setParam */
+ @Since("2.2.0")
+ def setMissingValue(value: Double): this.type = set(missingValue, value)
+
+ import org.apache.spark.ml.feature.Imputer._
+ setDefault(strategy -> mean, missingValue -> Double.NaN)
+
+ override def fit(dataset: Dataset[_]): ImputerModel = {
+ transformSchema(dataset.schema, logging = true)
+ val spark = dataset.sparkSession
+ import spark.implicits._
+ val surrogates = $(inputCols).map { inputCol =>
+ val ic = col(inputCol)
+ val filtered = dataset.select(ic.cast(DoubleType))
+ .filter(ic.isNotNull && ic =!= $(missingValue) && !ic.isNaN)
+ if(filtered.rdd.isEmpty()) {
+ throw new SparkException(s"surrogate cannot be computed. " +
+ s"All the values in $inputCol are Null, Nan or missingValue
($missingValue)")
+ }
+ val surrogate = $(strategy) match {
+ case Imputer.mean =>
filtered.select(avg(inputCol)).as[Double].first()
+ case Imputer.median => filtered.stat.approxQuantile(inputCol,
Array(0.5), 0.001).head
--- End diff --
Not really sure about the relative error here - perhaps `0.01` is
sufficient?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
