Github user tomerk commented on a diff in the pull request: https://github.com/apache/spark/pull/3637#discussion_r22752722 --- Diff: mllib/src/main/scala/org/apache/spark/ml/classification/ProbabilisticClassifier.scala --- @@ -0,0 +1,143 @@ +/* + * 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.annotation.{AlphaComponent, DeveloperApi} +import org.apache.spark.ml.param.{HasProbabilityCol, ParamMap, Params} +import org.apache.spark.mllib.linalg.{Vector, VectorUDT} +import org.apache.spark.sql._ +import org.apache.spark.sql.catalyst.analysis.Star + +/** + * Params for probabilistic classification. + */ +private[classification] trait ProbabilisticClassifierParams + extends ClassifierParams with HasProbabilityCol { + + override protected def validateAndTransformSchema( + schema: StructType, + paramMap: ParamMap, + fitting: Boolean, + featuresDataType: DataType): StructType = { + val parentSchema = super.validateAndTransformSchema(schema, paramMap, fitting, featuresDataType) + val map = this.paramMap ++ paramMap + addOutputColumn(parentSchema, map(probabilityCol), new VectorUDT) + } +} + + +/** + * :: AlphaComponent :: + * + * Single-label binary or multiclass classifier which can output class conditional probabilities. + * + * @tparam FeaturesType Type of input features. E.g., [[Vector]] + * @tparam Learner Concrete Estimator type + * @tparam M Concrete Model type + */ +@AlphaComponent +abstract class ProbabilisticClassifier[ + FeaturesType, + Learner <: ProbabilisticClassifier[FeaturesType, Learner, M], + M <: ProbabilisticClassificationModel[FeaturesType, M]] + extends Classifier[FeaturesType, Learner, M] with ProbabilisticClassifierParams { + + def setProbabilityCol(value: String): Learner = set(probabilityCol, value).asInstanceOf[Learner] +} + + +/** + * :: AlphaComponent :: + * + * Model produced by a [[ProbabilisticClassifier]]. + * Classes are indexed {0, 1, ..., numClasses - 1}. + * + * @tparam FeaturesType Type of input features. E.g., [[Vector]] + * @tparam M Concrete Model type + */ +@AlphaComponent +abstract class ProbabilisticClassificationModel[ + FeaturesType, + M <: ProbabilisticClassificationModel[FeaturesType, M]] + extends ClassificationModel[FeaturesType, M] with ProbabilisticClassifierParams { + + def setProbabilityCol(value: String): M = set(probabilityCol, value).asInstanceOf[M] + + /** + * Transforms dataset by reading from [[featuresCol]], and appending new columns as specified by + * parameters: + * - predicted labels as [[predictionCol]] of type [[Double]] + * - raw predictions (confidences) as [[rawPredictionCol]] of type [[Vector]] + * - probability of each class as [[probabilityCol]] of type [[Vector]]. + * + * @param dataset input dataset + * @param paramMap additional parameters, overwrite embedded params + * @return transformed dataset + */ + override def transform(dataset: SchemaRDD, paramMap: ParamMap): SchemaRDD = { + // This default implementation should be overridden as needed. + import dataset.sqlContext._ + import org.apache.spark.sql.catalyst.dsl._ + + // Check schema + transformSchema(dataset.schema, paramMap, logging = true) + val map = this.paramMap ++ paramMap + + // Prepare model + val tmpModel = if (paramMap.size != 0) { + val tmpModel = this.copy() + Params.inheritValues(paramMap, parent, tmpModel) + tmpModel + } else { + this + } + + val (numColsOutput, outputData) = + ClassificationModel.transformColumnsImpl[FeaturesType](dataset, tmpModel, map) + + // Output selected columns only. + if (map(probabilityCol) != "") { + // output probabilities + val features2probs: FeaturesType => Vector = (features) => { --- End diff -- In order for ```features2probs.call(*)``` to do the implicit conversion to a UDF
--- 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