diff --git a/mllib/src/main/scala/org/apache/spark/ml/classification/DecisionTreeClassifier.scala b/mllib/src/main/scala/org/apache/spark/ml/classification/DecisionTreeClassifier.scala
index 9f60f0896ec52..a40c0e03674b5 100644
--- a/mllib/src/main/scala/org/apache/spark/ml/classification/DecisionTreeClassifier.scala
+++ b/mllib/src/main/scala/org/apache/spark/ml/classification/DecisionTreeClassifier.scala
@@ -76,6 +76,10 @@ class DecisionTreeClassifier @Since("1.4.0") (
   @Since("1.4.0")
   override def setCacheNodeIds(value: Boolean): this.type = set(cacheNodeIds, value)
 
+  /** @group expertSetParam */
+  @Since("2.2.0")
+  def setCanMergeChildren(value: Boolean): this.type = set(canMergeChildren, value)
+
   /**
    * Specifies how often to checkpoint the cached node IDs.
    * E.g. 10 means that the cache will get checkpointed every 10 iterations.
diff --git a/mllib/src/main/scala/org/apache/spark/ml/classification/GBTClassifier.scala b/mllib/src/main/scala/org/apache/spark/ml/classification/GBTClassifier.scala
index ade0960f87a0d..28e0b65e7bcad 100644
--- a/mllib/src/main/scala/org/apache/spark/ml/classification/GBTClassifier.scala
+++ b/mllib/src/main/scala/org/apache/spark/ml/classification/GBTClassifier.scala
@@ -92,6 +92,10 @@ class GBTClassifier @Since("1.4.0") (
   @Since("1.4.0")
   override def setCacheNodeIds(value: Boolean): this.type = set(cacheNodeIds, value)
 
+  /** @group expertSetParam */
+  @Since("2.2.0")
+  def setCanMergeChildren(value: Boolean): this.type = set(canMergeChildren, value)
+
   /**
    * Specifies how often to checkpoint the cached node IDs.
    * E.g. 10 means that the cache will get checkpointed every 10 iterations.
diff --git a/mllib/src/main/scala/org/apache/spark/ml/classification/RandomForestClassifier.scala b/mllib/src/main/scala/org/apache/spark/ml/classification/RandomForestClassifier.scala
index ab4c235209289..b9e1745f02d01 100644
--- a/mllib/src/main/scala/org/apache/spark/ml/classification/RandomForestClassifier.scala
+++ b/mllib/src/main/scala/org/apache/spark/ml/classification/RandomForestClassifier.scala
@@ -78,6 +78,10 @@ class RandomForestClassifier @Since("1.4.0") (
   @Since("1.4.0")
   override def setCacheNodeIds(value: Boolean): this.type = set(cacheNodeIds, value)
 
+  /** @group expertSetParam */
+  @Since("2.2.0")
+  def setCanMergeChildren(value: Boolean): this.type = set(canMergeChildren, value)
+
   /**
    * Specifies how often to checkpoint the cached node IDs.
    * E.g. 10 means that the cache will get checkpointed every 10 iterations.
diff --git a/mllib/src/main/scala/org/apache/spark/ml/tree/Node.scala b/mllib/src/main/scala/org/apache/spark/ml/tree/Node.scala
index 07e98a142b10e..278ba8d76ec0d 100644
--- a/mllib/src/main/scala/org/apache/spark/ml/tree/Node.scala
+++ b/mllib/src/main/scala/org/apache/spark/ml/tree/Node.scala
@@ -342,6 +342,53 @@ private[tree] object LearningNode {
     new LearningNode(nodeIndex, None, None, None, false, null)
   }
 
+  /** merge the pair of leave of same parent recursively if their prediction is same.  */
+  def mergeChildrenWithSamePrediction(node: LearningNode): Int = {
+
+    /** find the mergeable leave and merge them. */
+    def checkAndMerge(node: LearningNode): (Option[Double], Int) = {
+      if (node == null) {
+        (None, 0)
+
+      } else if (node.isLeaf ||
+                 // sometimes, node is terminal while its isLeaf is not set.
+                 (node.leftChild.isEmpty && node.rightChild.isEmpty)) {
+        (Some(node.stats.impurityCalculator.predict), 0)
+
+      } else {
+        val (leftNode, leftMergeCounts) = checkAndMerge(node.leftChild.orNull)
+        val (rightNode, rightMergeCounts) = checkAndMerge(node.rightChild.orNull)
+        val mergeCounts = leftMergeCounts + rightMergeCounts
+
+        if (leftNode.isDefined && rightNode.isDefined && leftNode == rightNode) {
+          removeChildren(node)
+          node.isLeaf = true
+
+          (Some(node.stats.impurityCalculator.predict), mergeCounts + 1)
+
+        } else {
+          (None, mergeCounts)
+        }
+      }
+    }
+
+    val (_, mergeCounts) = checkAndMerge(node)
+
+    mergeCounts
+  }
+
+  /** delete all children of one node. */
+  def removeChildren(learningNode: LearningNode): Unit = {
+    val left = learningNode.leftChild
+    val right = learningNode.rightChild
+
+    learningNode.leftChild = None
+    learningNode.rightChild = None
+
+    left.foreach(removeChildren)
+    right.foreach(removeChildren)
+  }
+
   // The below indexing methods were copied from spark.mllib.tree.model.Node
 
   /**
diff --git a/mllib/src/main/scala/org/apache/spark/ml/tree/impl/RandomForest.scala b/mllib/src/main/scala/org/apache/spark/ml/tree/impl/RandomForest.scala
index 82e1ed85a0a14..9b53237a3a73d 100644
--- a/mllib/src/main/scala/org/apache/spark/ml/tree/impl/RandomForest.scala
+++ b/mllib/src/main/scala/org/apache/spark/ml/tree/impl/RandomForest.scala
@@ -217,6 +217,23 @@ private[spark] object RandomForest extends Logging {
       }
     }
 
+    // prune tree.
+    if (strategy.canMergeChildren) {
+      logInfo("Merge children with same prediction:")
+
+      // merge the pair of leaves if possible.
+      val mergeCounts = topNodes.map(LearningNode.mergeChildrenWithSamePrediction)
+
+      val mergeCountsOfTreesInfo: String = mergeCounts
+        .zipWithIndex
+        .map(_.swap)
+        .filter(_._2 > 0)
+        .map { case (id, count) => s"tree: $id, num of nodes merged: $count" }
+        .mkString("Merge info:\n", "\n", "\n")
+
+      logInfo(mergeCountsOfTreesInfo)
+    }
+
     val numFeatures = metadata.numFeatures
 
     parentUID match {
diff --git a/mllib/src/main/scala/org/apache/spark/ml/tree/treeParams.scala b/mllib/src/main/scala/org/apache/spark/ml/tree/treeParams.scala
index 3fc3ac58b7795..c4e8c48a6b31e 100644
--- a/mllib/src/main/scala/org/apache/spark/ml/tree/treeParams.scala
+++ b/mllib/src/main/scala/org/apache/spark/ml/tree/treeParams.scala
@@ -106,8 +106,18 @@ private[ml] trait DecisionTreeParams extends PredictorParams
     " algorithm will cache node IDs for each instance. Caching can speed up training of deeper" +
     " trees.")
 
+  /**
+   * If true, tree will try to merge its leaf nodes which share the same parent and output
+   * the same prediction.
+   * (default = false)
+   * @group expertParam
+   */
+  final val canMergeChildren: BooleanParam = new BooleanParam(this, "canMergeChildren", "If true," +
+    " the tree will try to merge its leaf nodes which share the same parent and output.")
+
   setDefault(maxDepth -> 5, maxBins -> 32, minInstancesPerNode -> 1, minInfoGain -> 0.0,
-    maxMemoryInMB -> 256, cacheNodeIds -> false, checkpointInterval -> 10)
+    maxMemoryInMB -> 256, cacheNodeIds -> false, canMergeChildren -> false,
+    checkpointInterval -> 10)
 
   /**
    * @deprecated This method is deprecated and will be removed in 3.0.0.
@@ -176,6 +186,9 @@ private[ml] trait DecisionTreeParams extends PredictorParams
   /** @group expertGetParam */
   final def getCacheNodeIds: Boolean = $(cacheNodeIds)
 
+  /** @group expertGetParam */
+  final def getCanMergeChildren: Boolean = $(canMergeChildren)
+
   /**
    * @deprecated This method is deprecated and will be removed in 3.0.0.
    * @group setParam
@@ -199,6 +212,7 @@ private[ml] trait DecisionTreeParams extends PredictorParams
     strategy.minInfoGain = getMinInfoGain
     strategy.minInstancesPerNode = getMinInstancesPerNode
     strategy.useNodeIdCache = getCacheNodeIds
+    strategy.canMergeChildren = getCanMergeChildren
     strategy.numClasses = numClasses
     strategy.categoricalFeaturesInfo = categoricalFeatures
     strategy.subsamplingRate = subsamplingRate
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 58e8f5be7b9f0..b584ae3a34c2d 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
@@ -61,6 +61,8 @@ import org.apache.spark.mllib.tree.impurity.{Entropy, Gini, Impurity, Variance}
  * @param subsamplingRate Fraction of the training data used for learning decision tree.
  * @param useNodeIdCache If this is true, instead of passing trees to executors, the algorithm will
  *                       maintain a separate RDD of node Id cache for each row.
+ * @param canMergeChildren Merge pairs of leaf nodes of the same parent which
+ *                         output the same prediction.
  * @param checkpointInterval How often to checkpoint when the node Id cache gets updated.
  *                           E.g. 10 means that the cache will get checkpointed every 10 updates. If
  *                           the checkpoint directory is not set in
@@ -80,6 +82,7 @@ class Strategy @Since("1.3.0") (
     @Since("1.0.0") @BeanProperty var maxMemoryInMB: Int = 256,
     @Since("1.2.0") @BeanProperty var subsamplingRate: Double = 1,
     @Since("1.2.0") @BeanProperty var useNodeIdCache: Boolean = false,
+    @Since("2.2.0") @BeanProperty var canMergeChildren: Boolean = false,
     @Since("1.2.0") @BeanProperty var checkpointInterval: Int = 10) extends Serializable {
 
   /**
@@ -172,7 +175,7 @@ class Strategy @Since("1.3.0") (
   def copy: Strategy = {
     new Strategy(algo, impurity, maxDepth, numClasses, maxBins,
       quantileCalculationStrategy, categoricalFeaturesInfo, minInstancesPerNode, minInfoGain,
-      maxMemoryInMB, subsamplingRate, useNodeIdCache, checkpointInterval)
+      maxMemoryInMB, subsamplingRate, useNodeIdCache, canMergeChildren, checkpointInterval)
   }
 }
 
diff --git a/mllib/src/test/scala/org/apache/spark/ml/classification/DecisionTreeClassifierSuite.scala b/mllib/src/test/scala/org/apache/spark/ml/classification/DecisionTreeClassifierSuite.scala
index 918ab27e2730b..2eb12c6fe0604 100644
--- a/mllib/src/test/scala/org/apache/spark/ml/classification/DecisionTreeClassifierSuite.scala
+++ b/mllib/src/test/scala/org/apache/spark/ml/classification/DecisionTreeClassifierSuite.scala
@@ -21,7 +21,7 @@ import org.apache.spark.SparkFunSuite
 import org.apache.spark.ml.feature.LabeledPoint
 import org.apache.spark.ml.linalg.{Vector, Vectors}
 import org.apache.spark.ml.param.ParamsSuite
-import org.apache.spark.ml.tree.{CategoricalSplit, InternalNode, LeafNode}
+import org.apache.spark.ml.tree.{CategoricalSplit, InternalNode, LeafNode, Node}
 import org.apache.spark.ml.tree.impl.TreeTests
 import org.apache.spark.ml.util.{DefaultReadWriteTest, MLTestingUtils}
 import org.apache.spark.mllib.regression.{LabeledPoint => OldLabeledPoint}
@@ -350,6 +350,31 @@ class DecisionTreeClassifierSuite
     dt.fit(df)
   }
 
+  test("SPARK-3159: Check for reducible DecisionTree") {
+    import DecisionTreeClassifierSuite.hasPairsOfSameChildren
+
+    val df: DataFrame = TreeTests.getIrisDataset(sc).toDF()
+    val dt = new DecisionTreeClassifier()
+      .setImpurity("Gini")
+      .setMaxBins(2)
+      .setMinInfoGain(0)
+      .setMinInstancesPerNode(1)
+      .setMaxDepth(3)
+      .setSeed(0)
+
+    val m1 = dt.setCanMergeChildren(false).fit(df)
+    assert(true === hasPairsOfSameChildren(m1.rootNode))
+
+    val m2 = dt.setCanMergeChildren(true).fit(df)
+    assert(false === hasPairsOfSameChildren(m2.rootNode))
+
+    val p1 = m1.transform(df).select(m1.getPredictionCol).collect()
+    val p2 = m2.transform(df).select(m2.getPredictionCol).collect()
+    assert(p1.zip(p2).forall { case (y1, y2) =>
+      y1.getDouble(0) === y2.getDouble(0)
+    })
+  }
+
   /////////////////////////////////////////////////////////////////////////////
   // Tests of model save/load
   /////////////////////////////////////////////////////////////////////////////
@@ -422,4 +447,25 @@ private[ml] object DecisionTreeClassifierSuite extends SparkFunSuite {
     TreeTests.checkEqual(oldTreeAsNew, newTree)
     assert(newTree.numFeatures === numFeatures)
   }
+
+  /** check if there exists pairs of leaf nodes with same prediction of the same parent. */
+  def hasPairsOfSameChildren(node: Node): Boolean = {
+    def check(node: Node): (Boolean, Option[Double]) = node match {
+      case n: LeafNode => (false, Some(n.prediction))
+      case n: InternalNode =>
+        val (leftFound, leftPredict) = check(n.leftChild)
+        val (rightFound, rightPredict) = check(n.rightChild)
+
+        if (leftFound || rightFound ||
+          (leftPredict.isDefined && leftPredict == rightPredict)) {
+          (true, None)
+        } else {
+          (false, None)
+        }
+    }
+
+    val (found, _) = check(node)
+
+    found
+  }
 }
diff --git a/mllib/src/test/scala/org/apache/spark/ml/tree/LearningNodeSuite.scala b/mllib/src/test/scala/org/apache/spark/ml/tree/LearningNodeSuite.scala
new file mode 100644
index 0000000000000..739b343e805d8
--- /dev/null
+++ b/mllib/src/test/scala/org/apache/spark/ml/tree/LearningNodeSuite.scala
@@ -0,0 +1,178 @@
+/*
+ * 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.tree
+
+import org.apache.spark.SparkFunSuite
+import org.apache.spark.mllib.tree.impurity.GiniCalculator
+import org.apache.spark.mllib.tree.model.ImpurityStats
+import org.apache.spark.mllib.util.MLlibTestSparkContext
+
+/**
+ * Test suite for [[LearningNode]].
+ */
+class LearningNodeSuite extends SparkFunSuite with MLlibTestSparkContext {
+  import LearningNodeSuite._
+
+  test("SPARK-3159: Check for reducible DecisionTree") {
+    val classA = TreeUtils.makeImpurityStats(2, 0)
+    val classB = TreeUtils.makeImpurityStats(2, 1)
+
+    val root = TreeUtils.fillBinaryTree(classA)(3)
+    NodeUtils.setAsLeaf(root, Array(5))
+    NodeUtils.setPredictValue(root, Array(6, 7, 12, 13, 14), classB)
+    /**
+     * use TreeUtils.toDebugTreeString(root).
+     * input:
+                8(0.0)
+             4--|
+                9(0.0)
+          2--|
+             5(0.0)
+       1--|
+                12(1.0)
+             6--|
+                13(1.0)
+          3--|
+                14(1.0)
+             7--|
+                15(0.0)
+      */
+    assert(true === hasPairsOfSameChildren(root))
+
+    val mergeCounts = LearningNode.mergeChildrenWithSamePrediction(root)
+    /**
+     * use TreeUtils.toDebugTreeString(root).
+     * result:
+          2(0.0)
+       1--|
+             6(1.0)
+          3--|
+                14(1.0)
+             7--|
+                15(0.0)
+      */
+    assert(false === hasPairsOfSameChildren(root))
+    assert(mergeCounts === 3)
+  }
+}
+
+object LearningNodeSuite {
+
+  /** check if there exists pairs of leaf nodes with same prediction of the same parent. */
+  def hasPairsOfSameChildren(node: LearningNode): Boolean =
+    if (node.isLeaf) {
+      false
+
+    } else {
+      val left = node.leftChild.get
+      val right = node.rightChild.get
+
+      if (left.isLeaf && right.isLeaf) {
+        val leftPredict = left.stats.impurityCalculator.predict
+        val rightPredict = right.stats.impurityCalculator.predict
+
+        leftPredict == rightPredict
+
+      } else {
+        // shortcut if find.
+        hasPairsOfSameChildren(left) || hasPairsOfSameChildren(right)
+      }
+    }
+
+  /** helper methods for constructing tree. */
+  object TreeUtils {
+    import LearningNode._
+
+    /** construct a full binary tree with same impurityStats. */
+    def fillBinaryTree(impurityStats: ImpurityStats)(maxHeight: Int): LearningNode = {
+      def create(id: Int, height: Int): LearningNode = {
+        if (height == 0) {
+          new LearningNode(id, None, None, None, true, impurityStats)
+
+        } else {
+          val leftNode = create(leftChildIndex(id), height - 1)
+          val rightNode = create(rightChildIndex(id), height - 1)
+          // use id to help locate node when debug.
+          val split = new ContinuousSplit(id, id)
+
+          new LearningNode(
+            id, Some(leftNode), Some(rightNode),
+            Some(split), false, impurityStats)
+        }
+      }
+
+      create(1, maxHeight)
+    }
+
+    /** create an ImpurityStats for classification. */
+    def makeImpurityStats(numClass: Int, predictClassId: Int): ImpurityStats = {
+      val stat = Array.fill(numClass)(0.0)
+      stat(predictClassId) = 1.0
+
+      val calculator = new GiniCalculator(stat)
+
+      new ImpurityStats(0.0, 0.0, calculator, null, null)
+    }
+
+    /** Full description of model */
+    def toDebugTreeString(node: LearningNode, indent: Int = 0): String = {
+      val prefix: String = "   " * indent
+      val id = node.id
+
+      if (node.isLeaf) {
+        val predict = node.stats.impurityCalculator.predict
+
+        prefix + id + "(" + predict + ")" + "\n"
+
+      } else {
+        val left = toDebugTreeString(node.leftChild.get, indent + 1)
+        val right = toDebugTreeString(node.rightChild.get, indent + 1)
+
+        left +
+        prefix + id + "--|\n" +
+        right
+      }
+    }
+  }
+
+  /** helper methods used to operate nodes. */
+  object NodeUtils {
+    import LearningNode._
+
+    /** assign the ImpurityStats to all nodes required in nodeIds. */
+    def setPredictValue(root: LearningNode,
+                        nodeIds: Array[Int],
+                        impurityStats: ImpurityStats): Unit =
+      nodeIds.foreach { id =>
+        val node = getNode(id, root)
+
+        node.stats = impurityStats
+      }
+
+    /** set internal nodes as leaf. */
+    def setAsLeaf(root: LearningNode, nodeIds: Array[Int]): Unit =
+      nodeIds.foreach { id =>
+        val node = getNode(id, root)
+
+        if (! node.isLeaf) {
+          removeChildren(node)
+          node.isLeaf = true
+        }
+      }
+  }
+}
diff --git a/mllib/src/test/scala/org/apache/spark/ml/tree/impl/TreeTests.scala b/mllib/src/test/scala/org/apache/spark/ml/tree/impl/TreeTests.scala
index 92a236928e90b..04e5d54ebd474 100644
--- a/mllib/src/test/scala/org/apache/spark/ml/tree/impl/TreeTests.scala
+++ b/mllib/src/test/scala/org/apache/spark/ml/tree/impl/TreeTests.scala
@@ -227,4 +227,163 @@ private[ml] object TreeTests extends SparkFunSuite {
       LabeledPoint(1.0, Vectors.dense(1.0, 2.0)))
     sc.parallelize(arr)
   }
+
+  /** Iris dataset.
+   * @see http://www.csie.ntu.edu.tw/~cjlin/libsvmtools/datasets/multiclass/iris.scale
+   */
+  def getIrisDataset(sc: SparkContext): RDD[LabeledPoint] = {
+    val arr = Array(
+      LabeledPoint(1, Vectors.dense(-0.555556, 0.25, -0.864407, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.666667, -0.166667, -0.864407, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.777778, 0.0, -0.898305, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.833333, -0.0833334, -0.830508, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.611111, 0.333333, -0.864407, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.388889, 0.583333, -0.762712, -0.75)),
+      LabeledPoint(1, Vectors.dense(-0.833333, 0.166667, -0.864407, -0.833333)),
+      LabeledPoint(1, Vectors.dense(-0.611111, 0.166667, -0.830508, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.944444, -0.25, -0.864407, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.666667, -0.0833334, -0.830508, -1.0)),
+      LabeledPoint(1, Vectors.dense(-0.388889, 0.416667, -0.830508, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.722222, 0.166667, -0.79661, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.722222, -0.166667, -0.864407, -1.0)),
+      LabeledPoint(1, Vectors.dense(-1.0, -0.166667, -0.966102, -1.0)),
+      LabeledPoint(1, Vectors.dense(-0.166667, 0.666667, -0.932203, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.222222, 1.0, -0.830508, -0.75)),
+      LabeledPoint(1, Vectors.dense(-0.388889, 0.583333, -0.898305, -0.75)),
+      LabeledPoint(1, Vectors.dense(-0.555556, 0.25, -0.864407, -0.833333)),
+      LabeledPoint(1, Vectors.dense(-0.222222, 0.5, -0.762712, -0.833333)),
+      LabeledPoint(1, Vectors.dense(-0.555556, 0.5, -0.830508, -0.833333)),
+      LabeledPoint(1, Vectors.dense(-0.388889, 0.166667, -0.762712, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.555556, 0.416667, -0.830508, -0.75)),
+      LabeledPoint(1, Vectors.dense(-0.833333, 0.333333, -1.0, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.555556, 0.0833333, -0.762712, -0.666667)),
+      LabeledPoint(1, Vectors.dense(-0.722222, 0.166667, -0.694915, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.611111, -0.166667, -0.79661, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.611111, 0.166667, -0.79661, -0.75)),
+      LabeledPoint(1, Vectors.dense(-0.5, 0.25, -0.830508, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.5, 0.166667, -0.864407, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.777778, 0.0, -0.79661, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.722222, -0.0833334, -0.79661, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.388889, 0.166667, -0.830508, -0.75)),
+      LabeledPoint(1, Vectors.dense(-0.5, 0.75, -0.830508, -1.0)),
+      LabeledPoint(1, Vectors.dense(-0.333333, 0.833333, -0.864407, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.666667, -0.0833334, -0.830508, -1.0)),
+      LabeledPoint(1, Vectors.dense(-0.611111, 0.0, -0.932203, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.333333, 0.25, -0.898305, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.666667, -0.0833334, -0.830508, -1.0)),
+      LabeledPoint(1, Vectors.dense(-0.944444, -0.166667, -0.898305, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.555556, 0.166667, -0.830508, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.611111, 0.25, -0.898305, -0.833333)),
+      LabeledPoint(1, Vectors.dense(-0.888889, -0.75, -0.898305, -0.833333)),
+      LabeledPoint(1, Vectors.dense(-0.944444, 0.0, -0.898305, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.611111, 0.25, -0.79661, -0.583333)),
+      LabeledPoint(1, Vectors.dense(-0.555556, 0.5, -0.694915, -0.75)),
+      LabeledPoint(1, Vectors.dense(-0.722222, -0.166667, -0.864407, -0.833333)),
+      LabeledPoint(1, Vectors.dense(-0.555556, 0.5, -0.79661, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.833333, 0.0, -0.864407, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.444444, 0.416667, -0.830508, -0.916667)),
+      LabeledPoint(1, Vectors.dense(-0.611111, 0.0833333, -0.864407, -0.916667)),
+      LabeledPoint(2, Vectors.dense(0.5, 0.0, 0.254237, 0.0833333)),
+      LabeledPoint(2, Vectors.dense(0.166667, 0.0, 0.186441, 0.166667)),
+      LabeledPoint(2, Vectors.dense(0.444444, -0.0833334, 0.322034, 0.166667)),
+      LabeledPoint(2, Vectors.dense(-0.333333, -0.75, 0.0169491, -4.03573E-8)),
+      LabeledPoint(2, Vectors.dense(0.222222, -0.333333, 0.220339, 0.166667)),
+      LabeledPoint(2, Vectors.dense(-0.222222, -0.333333, 0.186441, -4.03573E-8)),
+      LabeledPoint(2, Vectors.dense(0.111111, 0.0833333, 0.254237, 0.25)),
+      LabeledPoint(2, Vectors.dense(-0.666667, -0.666667, -0.220339, -0.25)),
+      LabeledPoint(2, Vectors.dense(0.277778, -0.25, 0.220339, -4.03573E-8)),
+      LabeledPoint(2, Vectors.dense(-0.5, -0.416667, -0.0169491, 0.0833333)),
+      LabeledPoint(2, Vectors.dense(-0.611111, -1.0, -0.152542, -0.25)),
+      LabeledPoint(2, Vectors.dense(-0.111111, -0.166667, 0.0847457, 0.166667)),
+      LabeledPoint(2, Vectors.dense(-0.0555556, -0.833333, 0.0169491, -0.25)),
+      LabeledPoint(2, Vectors.dense(-1.32455E-7, -0.25, 0.254237, 0.0833333)),
+      LabeledPoint(2, Vectors.dense(-0.277778, -0.25, -0.118644, -4.03573E-8)),
+      LabeledPoint(2, Vectors.dense(0.333333, -0.0833334, 0.152542, 0.0833333)),
+      LabeledPoint(2, Vectors.dense(-0.277778, -0.166667, 0.186441, 0.166667)),
+      LabeledPoint(2, Vectors.dense(-0.166667, -0.416667, 0.0508474, -0.25)),
+      LabeledPoint(2, Vectors.dense(0.0555554, -0.833333, 0.186441, 0.166667)),
+      LabeledPoint(2, Vectors.dense(-0.277778, -0.583333, -0.0169491, -0.166667)),
+      LabeledPoint(2, Vectors.dense(-0.111111, 0.0, 0.288136, 0.416667)),
+      LabeledPoint(2, Vectors.dense(-1.32455E-7, -0.333333, 0.0169491, -4.03573E-8)),
+      LabeledPoint(2, Vectors.dense(0.111111, -0.583333, 0.322034, 0.166667)),
+      LabeledPoint(2, Vectors.dense(-1.32455E-7, -0.333333, 0.254237, -0.0833333)),
+      LabeledPoint(2, Vectors.dense(0.166667, -0.25, 0.118644, -4.03573E-8)),
+      LabeledPoint(2, Vectors.dense(0.277778, -0.166667, 0.152542, 0.0833333)),
+      LabeledPoint(2, Vectors.dense(0.388889, -0.333333, 0.288136, 0.0833333)),
+      LabeledPoint(2, Vectors.dense(0.333333, -0.166667, 0.355932, 0.333333)),
+      LabeledPoint(2, Vectors.dense(-0.0555556, -0.25, 0.186441, 0.166667)),
+      LabeledPoint(2, Vectors.dense(-0.222222, -0.5, -0.152542, -0.25)),
+      LabeledPoint(2, Vectors.dense(-0.333333, -0.666667, -0.0508475, -0.166667)),
+      LabeledPoint(2, Vectors.dense(-0.333333, -0.666667, -0.0847458, -0.25)),
+      LabeledPoint(2, Vectors.dense(-0.166667, -0.416667, -0.0169491, -0.0833333)),
+      LabeledPoint(2, Vectors.dense(-0.0555556, -0.416667, 0.38983, 0.25)),
+      LabeledPoint(2, Vectors.dense(-0.388889, -0.166667, 0.186441, 0.166667)),
+      LabeledPoint(2, Vectors.dense(-0.0555556, 0.166667, 0.186441, 0.25)),
+      LabeledPoint(2, Vectors.dense(0.333333, -0.0833334, 0.254237, 0.166667)),
+      LabeledPoint(2, Vectors.dense(0.111111, -0.75, 0.152542, -4.03573E-8)),
+      LabeledPoint(2, Vectors.dense(-0.277778, -0.166667, 0.0508474, -4.03573E-8)),
+      LabeledPoint(2, Vectors.dense(-0.333333, -0.583333, 0.0169491, -4.03573E-8)),
+      LabeledPoint(2, Vectors.dense(-0.333333, -0.5, 0.152542, -0.0833333)),
+      LabeledPoint(2, Vectors.dense(-1.32455E-7, -0.166667, 0.220339, 0.0833333)),
+      LabeledPoint(2, Vectors.dense(-0.166667, -0.5, 0.0169491, -0.0833333)),
+      LabeledPoint(2, Vectors.dense(-0.611111, -0.75, -0.220339, -0.25)),
+      LabeledPoint(2, Vectors.dense(-0.277778, -0.416667, 0.0847457, -4.03573E-8)),
+      LabeledPoint(2, Vectors.dense(-0.222222, -0.166667, 0.0847457, -0.0833333)),
+      LabeledPoint(2, Vectors.dense(-0.222222, -0.25, 0.0847457, -4.03573E-8)),
+      LabeledPoint(2, Vectors.dense(0.0555554, -0.25, 0.118644, -4.03573E-8)),
+      LabeledPoint(2, Vectors.dense(-0.555556, -0.583333, -0.322034, -0.166667)),
+      LabeledPoint(2, Vectors.dense(-0.222222, -0.333333, 0.0508474, -4.03573E-8)),
+      LabeledPoint(3, Vectors.dense(0.111111, 0.0833333, 0.694915, 1.0)),
+      LabeledPoint(3, Vectors.dense(-0.166667, -0.416667, 0.38983, 0.5)),
+      LabeledPoint(3, Vectors.dense(0.555555, -0.166667, 0.661017, 0.666667)),
+      LabeledPoint(3, Vectors.dense(0.111111, -0.25, 0.559322, 0.416667)),
+      LabeledPoint(3, Vectors.dense(0.222222, -0.166667, 0.627119, 0.75)),
+      LabeledPoint(3, Vectors.dense(0.833333, -0.166667, 0.898305, 0.666667)),
+      LabeledPoint(3, Vectors.dense(-0.666667, -0.583333, 0.186441, 0.333333)),
+      LabeledPoint(3, Vectors.dense(0.666667, -0.25, 0.79661, 0.416667)),
+      LabeledPoint(3, Vectors.dense(0.333333, -0.583333, 0.627119, 0.416667)),
+      LabeledPoint(3, Vectors.dense(0.611111, 0.333333, 0.728813, 1.0)),
+      LabeledPoint(3, Vectors.dense(0.222222, 0.0, 0.38983, 0.583333)),
+      LabeledPoint(3, Vectors.dense(0.166667, -0.416667, 0.457627, 0.5)),
+      LabeledPoint(3, Vectors.dense(0.388889, -0.166667, 0.525424, 0.666667)),
+      LabeledPoint(3, Vectors.dense(-0.222222, -0.583333, 0.355932, 0.583333)),
+      LabeledPoint(3, Vectors.dense(-0.166667, -0.333333, 0.38983, 0.916667)),
+      LabeledPoint(3, Vectors.dense(0.166667, 0.0, 0.457627, 0.833333)),
+      LabeledPoint(3, Vectors.dense(0.222222, -0.166667, 0.525424, 0.416667)),
+      LabeledPoint(3, Vectors.dense(0.888889, 0.5, 0.932203, 0.75)),
+      LabeledPoint(3, Vectors.dense(0.888889, -0.5, 1.0, 0.833333)),
+      LabeledPoint(3, Vectors.dense(-0.0555556, -0.833333, 0.355932, 0.166667)),
+      LabeledPoint(3, Vectors.dense(0.444444, 0.0, 0.59322, 0.833333)),
+      LabeledPoint(3, Vectors.dense(-0.277778, -0.333333, 0.322034, 0.583333)),
+      LabeledPoint(3, Vectors.dense(0.888889, -0.333333, 0.932203, 0.583333)),
+      LabeledPoint(3, Vectors.dense(0.111111, -0.416667, 0.322034, 0.416667)),
+      LabeledPoint(3, Vectors.dense(0.333333, 0.0833333, 0.59322, 0.666667)),
+      LabeledPoint(3, Vectors.dense(0.611111, 0.0, 0.694915, 0.416667)),
+      LabeledPoint(3, Vectors.dense(0.0555554, -0.333333, 0.288136, 0.416667)),
+      LabeledPoint(3, Vectors.dense(-1.32455E-7, -0.166667, 0.322034, 0.416667)),
+      LabeledPoint(3, Vectors.dense(0.166667, -0.333333, 0.559322, 0.666667)),
+      LabeledPoint(3, Vectors.dense(0.611111, -0.166667, 0.627119, 0.25)),
+      LabeledPoint(3, Vectors.dense(0.722222, -0.333333, 0.728813, 0.5)),
+      LabeledPoint(3, Vectors.dense(1.0, 0.5, 0.830508, 0.583333)),
+      LabeledPoint(3, Vectors.dense(0.166667, -0.333333, 0.559322, 0.75)),
+      LabeledPoint(3, Vectors.dense(0.111111, -0.333333, 0.38983, 0.166667)),
+      LabeledPoint(3, Vectors.dense(-1.32455E-7, -0.5, 0.559322, 0.0833333)),
+      LabeledPoint(3, Vectors.dense(0.888889, -0.166667, 0.728813, 0.833333)),
+      LabeledPoint(3, Vectors.dense(0.111111, 0.166667, 0.559322, 0.916667)),
+      LabeledPoint(3, Vectors.dense(0.166667, -0.0833334, 0.525424, 0.416667)),
+      LabeledPoint(3, Vectors.dense(-0.0555556, -0.166667, 0.288136, 0.416667)),
+      LabeledPoint(3, Vectors.dense(0.444444, -0.0833334, 0.491525, 0.666667)),
+      LabeledPoint(3, Vectors.dense(0.333333, -0.0833334, 0.559322, 0.916667)),
+      LabeledPoint(3, Vectors.dense(0.444444, -0.0833334, 0.38983, 0.833333)),
+      LabeledPoint(3, Vectors.dense(-0.166667, -0.416667, 0.38983, 0.5)),
+      LabeledPoint(3, Vectors.dense(0.388889, 0.0, 0.661017, 0.833333)),
+      LabeledPoint(3, Vectors.dense(0.333333, 0.0833333, 0.59322, 1.0)),
+      LabeledPoint(3, Vectors.dense(0.333333, -0.166667, 0.423729, 0.833333)),
+      LabeledPoint(3, Vectors.dense(0.111111, -0.583333, 0.355932, 0.5)),
+      LabeledPoint(3, Vectors.dense(0.222222, -0.166667, 0.423729, 0.583333)),
+      LabeledPoint(3, Vectors.dense(0.0555554, 0.166667, 0.491525, 0.833333)),
+      LabeledPoint(3, Vectors.dense(-0.111111, -0.166667, 0.38983, 0.416667)))
+
+    sc.parallelize(arr)
+  }
 }
diff --git a/project/MimaExcludes.scala b/project/MimaExcludes.scala
index 1793da03a2c3e..e7c2add8acf52 100644
--- a/project/MimaExcludes.scala
+++ b/project/MimaExcludes.scala
@@ -36,6 +36,10 @@ object MimaExcludes {
 
   // Exclude rules for 2.3.x
   lazy val v23excludes = v22excludes ++ Seq(
+    // [SPARK-3159][MLlib] Check for reducible DecisionTree.
+    ProblemFilters.exclude[IncompatibleResultTypeProblem]("org.apache.spark.mllib.tree.configuration.Strategy.<init>$default$13")
+    ProblemFilters.exclude[DirectMissingMethodProblem]("org.apache.spark.mllib.tree.configuration.Strategy.this")
+
     // [SPARK-20495][SQL] Add StorageLevel to cacheTable API
     ProblemFilters.exclude[ReversedMissingMethodProblem]("org.apache.spark.sql.catalog.Catalog.cacheTable"),