diff --git a/mllib/src/main/scala/org/apache/spark/mllib/classification/NaiveBayes.scala b/mllib/src/main/scala/org/apache/spark/mllib/classification/NaiveBayes.scala index 37eaaa28a9619..2026d5ba5270c 100644 --- a/mllib/src/main/scala/org/apache/spark/mllib/classification/NaiveBayes.scala +++ b/mllib/src/main/scala/org/apache/spark/mllib/classification/NaiveBayes.scala @@ -41,7 +41,7 @@ class NaiveBayesModel( private val brzTheta = new BDM[Double](theta.length, theta(0).length) { - // Need to put an extra pair of braces to prevent Scala treat `i` as a member. + // Need to put an extra pair of braces to prevent Scala treating `i` as a member. var i = 0 while (i < theta.length) { var j = 0 diff --git a/mllib/src/main/scala/org/apache/spark/mllib/tree/DecisionTree.scala b/mllib/src/main/scala/org/apache/spark/mllib/tree/DecisionTree.scala index dee9594a9dd79..04e7e4241910e 100644 --- a/mllib/src/main/scala/org/apache/spark/mllib/tree/DecisionTree.scala +++ b/mllib/src/main/scala/org/apache/spark/mllib/tree/DecisionTree.scala @@ -33,13 +33,15 @@ import org.apache.spark.util.random.XORShiftRandom import org.apache.spark.mllib.linalg.{Vector, Vectors} /** + * EXPERIMENTAL + * * A class that implements a decision tree algorithm for classification and regression. It * supports both continuous and categorical features. * @param strategy The configuration parameters for the tree algorithm which specify the type * of algorithm (classification, regression, etc.), feature type (continuous, * categorical), depth of the tree, quantile calculation strategy, etc. */ -class DecisionTree private(val strategy: Strategy) extends Serializable with Logging { +class DecisionTree (private val strategy: Strategy) extends Serializable with Logging { /** * Method to train a decision tree model over an RDD diff --git a/mllib/src/main/scala/org/apache/spark/mllib/tree/configuration/Strategy.scala b/mllib/src/main/scala/org/apache/spark/mllib/tree/configuration/Strategy.scala index df565f3eb8859..0cbe7d73cddad 100644 --- a/mllib/src/main/scala/org/apache/spark/mllib/tree/configuration/Strategy.scala +++ b/mllib/src/main/scala/org/apache/spark/mllib/tree/configuration/Strategy.scala @@ -40,4 +40,4 @@ class Strategy ( val maxDepth: Int, val maxBins: Int = 100, val quantileCalculationStrategy: QuantileStrategy = Sort, - val categoricalFeaturesInfo: Map[Int,Int] = Map[Int,Int]()) extends Serializable + val categoricalFeaturesInfo: Map[Int, Int] = Map[Int, Int]()) extends Serializable diff --git a/mllib/src/main/scala/org/apache/spark/mllib/tree/impurity/Entropy.scala b/mllib/src/main/scala/org/apache/spark/mllib/tree/impurity/Entropy.scala index b93995fcf9441..beec48bb3a108 100644 --- a/mllib/src/main/scala/org/apache/spark/mllib/tree/impurity/Entropy.scala +++ b/mllib/src/main/scala/org/apache/spark/mllib/tree/impurity/Entropy.scala @@ -18,20 +18,24 @@ package org.apache.spark.mllib.tree.impurity /** + * EXPERIMENTAL + * * Class for calculating [[http://en.wikipedia.org/wiki/Binary_entropy_function entropy]] during * binary classification. */ object Entropy extends Impurity { - def log2(x: Double) = scala.math.log(x) / scala.math.log(2) + private[tree] def log2(x: Double) = scala.math.log(x) / scala.math.log(2) /** + * DEVELOPER API + * * entropy calculation * @param c0 count of instances with label 0 * @param c1 count of instances with label 1 * @return entropy value */ - def calculate(c0: Double, c1: Double): Double = { + override def calculate(c0: Double, c1: Double): Double = { if (c0 == 0 || c1 == 0) { 0 } else { @@ -42,6 +46,6 @@ object Entropy extends Impurity { } } - def calculate(count: Double, sum: Double, sumSquares: Double): Double = + override def calculate(count: Double, sum: Double, sumSquares: Double): Double = throw new UnsupportedOperationException("Entropy.calculate") } diff --git a/mllib/src/main/scala/org/apache/spark/mllib/tree/impurity/Gini.scala b/mllib/src/main/scala/org/apache/spark/mllib/tree/impurity/Gini.scala index c0407554a91b3..5babe7d10d111 100644 --- a/mllib/src/main/scala/org/apache/spark/mllib/tree/impurity/Gini.scala +++ b/mllib/src/main/scala/org/apache/spark/mllib/tree/impurity/Gini.scala @@ -18,6 +18,8 @@ package org.apache.spark.mllib.tree.impurity /** + * EXPERIMENTAL + * * Class for calculating the * [[http://en.wikipedia.org/wiki/Decision_tree_learning#Gini_impurity Gini impurity]] * during binary classification. @@ -25,6 +27,8 @@ package org.apache.spark.mllib.tree.impurity object Gini extends Impurity { /** + * DEVELOPER API + * * Gini coefficient calculation * @param c0 count of instances with label 0 * @param c1 count of instances with label 1 @@ -41,6 +45,6 @@ object Gini extends Impurity { } } - def calculate(count: Double, sum: Double, sumSquares: Double): Double = + override def calculate(count: Double, sum: Double, sumSquares: Double): Double = throw new UnsupportedOperationException("Gini.calculate") } diff --git a/mllib/src/main/scala/org/apache/spark/mllib/tree/impurity/Impurity.scala b/mllib/src/main/scala/org/apache/spark/mllib/tree/impurity/Impurity.scala index 43f296ac56bc8..e6fa115030e7a 100644 --- a/mllib/src/main/scala/org/apache/spark/mllib/tree/impurity/Impurity.scala +++ b/mllib/src/main/scala/org/apache/spark/mllib/tree/impurity/Impurity.scala @@ -18,12 +18,14 @@ package org.apache.spark.mllib.tree.impurity /** + * EXPERIMENTAL + * * Trait for calculating information gain. */ trait Impurity extends Serializable { /** - * DEVELOPER API - UNSTABLE + * DEVELOPER API * * information calculation for binary classification * @param c0 count of instances with label 0 @@ -33,7 +35,7 @@ trait Impurity extends Serializable { def calculate(c0 : Double, c1 : Double): Double /** - * DEVELOPER API - UNSTABLE + * DEVELOPER API * * information calculation for regression * @param count number of instances diff --git a/mllib/src/main/scala/org/apache/spark/mllib/tree/impurity/Variance.scala b/mllib/src/main/scala/org/apache/spark/mllib/tree/impurity/Variance.scala index b74577dcec167..7be3b9236ecd9 100644 --- a/mllib/src/main/scala/org/apache/spark/mllib/tree/impurity/Variance.scala +++ b/mllib/src/main/scala/org/apache/spark/mllib/tree/impurity/Variance.scala @@ -18,6 +18,8 @@ package org.apache.spark.mllib.tree.impurity /** + * EXPERIMENTAL + * * Class for calculating variance during regression */ object Variance extends Impurity { @@ -25,6 +27,8 @@ object Variance extends Impurity { throw new UnsupportedOperationException("Variance.calculate") /** + * DEVELOPER API + * * variance calculation * @param count number of instances * @param sum sum of labels diff --git a/mllib/src/main/scala/org/apache/spark/mllib/tree/model/Bin.scala b/mllib/src/main/scala/org/apache/spark/mllib/tree/model/Bin.scala index a57faa13745f7..2d71e1e366069 100644 --- a/mllib/src/main/scala/org/apache/spark/mllib/tree/model/Bin.scala +++ b/mllib/src/main/scala/org/apache/spark/mllib/tree/model/Bin.scala @@ -30,4 +30,5 @@ import org.apache.spark.mllib.tree.configuration.FeatureType._ * @param featureType type of feature -- categorical or continuous * @param category categorical label value accepted in the bin */ +private[tree] case class Bin(lowSplit: Split, highSplit: Split, featureType: FeatureType, category: Double) diff --git a/mllib/src/main/scala/org/apache/spark/mllib/tree/model/DecisionTreeModel.scala b/mllib/src/main/scala/org/apache/spark/mllib/tree/model/DecisionTreeModel.scala index a6dca84a2ce09..e336ea74e3b76 100644 --- a/mllib/src/main/scala/org/apache/spark/mllib/tree/model/DecisionTreeModel.scala +++ b/mllib/src/main/scala/org/apache/spark/mllib/tree/model/DecisionTreeModel.scala @@ -22,6 +22,8 @@ import org.apache.spark.rdd.RDD import org.apache.spark.mllib.linalg.Vector /** + * EXPERIMENTAL + * * Model to store the decision tree parameters * @param topNode root node * @param algo algorithm type -- classification or regression diff --git a/mllib/src/main/scala/org/apache/spark/mllib/tree/model/Filter.scala b/mllib/src/main/scala/org/apache/spark/mllib/tree/model/Filter.scala index ebc9595eafef3..2deaf4ae8dcab 100644 --- a/mllib/src/main/scala/org/apache/spark/mllib/tree/model/Filter.scala +++ b/mllib/src/main/scala/org/apache/spark/mllib/tree/model/Filter.scala @@ -22,7 +22,7 @@ package org.apache.spark.mllib.tree.model * @param split split specifying the feature index, type and threshold * @param comparison integer specifying <,=,> */ -case class Filter(split: Split, comparison: Int) { +private[tree] case class Filter(split: Split, comparison: Int) { // Comparison -1,0,1 signifies <.=,> override def toString = " split = " + split + "comparison = " + comparison } diff --git a/mllib/src/main/scala/org/apache/spark/mllib/tree/model/InformationGainStats.scala b/mllib/src/main/scala/org/apache/spark/mllib/tree/model/InformationGainStats.scala index 99bf79cf12e45..aa1a478ea41b5 100644 --- a/mllib/src/main/scala/org/apache/spark/mllib/tree/model/InformationGainStats.scala +++ b/mllib/src/main/scala/org/apache/spark/mllib/tree/model/InformationGainStats.scala @@ -18,6 +18,8 @@ package org.apache.spark.mllib.tree.model /** + * DEVELOPER API + * * Information gain statistics for each split * @param gain information gain value * @param impurity current node impurity diff --git a/mllib/src/main/scala/org/apache/spark/mllib/tree/model/Node.scala b/mllib/src/main/scala/org/apache/spark/mllib/tree/model/Node.scala index aac3f9ce308f7..361361f937c76 100644 --- a/mllib/src/main/scala/org/apache/spark/mllib/tree/model/Node.scala +++ b/mllib/src/main/scala/org/apache/spark/mllib/tree/model/Node.scala @@ -22,6 +22,8 @@ import org.apache.spark.mllib.tree.configuration.FeatureType._ import org.apache.spark.mllib.linalg.Vector /** + * DEVELOPER API + * * Node in a decision tree * @param id integer node id * @param predict predicted value at the node diff --git a/mllib/src/main/scala/org/apache/spark/mllib/tree/model/Split.scala b/mllib/src/main/scala/org/apache/spark/mllib/tree/model/Split.scala index 4e64a81dda74e..1ceb64ca44290 100644 --- a/mllib/src/main/scala/org/apache/spark/mllib/tree/model/Split.scala +++ b/mllib/src/main/scala/org/apache/spark/mllib/tree/model/Split.scala @@ -30,7 +30,7 @@ case class Split( feature: Int, threshold: Double, featureType: FeatureType, - categories: List[Double]){ + categories: List[Double]) { override def toString = "Feature = " + feature + ", threshold = " + threshold + ", featureType = " + featureType + @@ -42,7 +42,7 @@ case class Split( * @param feature feature index * @param featureType type of feature -- categorical or continuous */ -class DummyLowSplit(feature: Int, featureType: FeatureType) +private[tree] class DummyLowSplit(feature: Int, featureType: FeatureType) extends Split(feature, Double.MinValue, featureType, List()) /** @@ -50,7 +50,7 @@ class DummyLowSplit(feature: Int, featureType: FeatureType) * @param feature feature index * @param featureType type of feature -- categorical or continuous */ -class DummyHighSplit(feature: Int, featureType: FeatureType) +private[tree] class DummyHighSplit(feature: Int, featureType: FeatureType) extends Split(feature, Double.MaxValue, featureType, List()) /** @@ -59,6 +59,6 @@ class DummyHighSplit(feature: Int, featureType: FeatureType) * @param feature feature index * @param featureType type of feature -- categorical or continuous */ -class DummyCategoricalSplit(feature: Int, featureType: FeatureType) +private[tree] class DummyCategoricalSplit(feature: Int, featureType: FeatureType) extends Split(feature, Double.MaxValue, featureType, List())