Get all leaves of the DecisionTree ( then construct spline thru leafnodes to build f(x)=>y )

result.txt

// Exiting paste mode, now interpreting.

id = 8, isLeaf = true, predict = 0.0 (prob = -1.0), impurity = 0.0, split = None, stats = None
id = 9, isLeaf = true, predict = 1.4736842105263157 (prob = -1.0), impurity = 0.2493074792243767, split = None, stats = None
id = 10, isLeaf = true, predict = 3.0 (prob = -1.0), impurity = 0.16666666666666666, split = None, stats = None
id = 11, isLeaf = true, predict = 4.1 (prob = -1.0), impurity = 0.09000000000000057, split = None, stats = None
id = 12, isLeaf = true, predict = 5.0 (prob = -1.0), impurity = 0.0, split = None, stats = None
id = 13, isLeaf = true, predict = 6.444444444444445 (prob = -1.0), impurity = 0.2469135802469143, split = None, stats = None
id = 14, isLeaf = true, predict = 7.923076923076923 (prob = -1.0), impurity = 0.2248520710059158, split = None, stats = None
id = 15, isLeaf = true, predict = 9.0 (prob = -1.0), impurity = 0.0, split = None, stats = None

tree.scala

import org.apache.spark.mllib.regression._
import org.apache.spark.mllib.linalg._
import org.apache.spark.mllib.tree.configuration._
import org.apache.spark.mllib.tree._
import org.apache.spark.mllib.tree.model.Node

def getLeaves(nl:List[Node]):List[Node] = {
    var res = nl.map{ n=>
    val l:List[Node] = if (n.leftNode.isDefined) List(n.leftNode.get) else List()
    val r:List[Node] = if (n.rightNode.isDefined) List(n.rightNode.get) else List()
    l ++ r ++ List(n)
  }.flatten
  .distinct
  if (res.size != nl.size) res = getLeaves(res)
  res.filter{ x=> x.isLeaf }
}

val data = (1 until 100).toList
.map{ x=> (x,x/10) }
.map{ case (feature:Int, label:Int) => 
  new LabeledPoint(label, new DenseVector(Array(feature))) 
}

val strategy = Strategy.defaultStrategy(Algo.Regression)
strategy.setMaxDepth(3)
val tree = new DecisionTree(strategy)
val model = tree.run(sc.makeRDD(data))
getLeaves(List(root)).foreach(println)