Github user srowen commented on a diff in the pull request:
https://github.com/apache/spark/pull/13624#discussion_r68205142
--- Diff:
mllib/src/main/scala/org/apache/spark/ml/tree/impl/GradientBoostedTrees.scala
---
@@ -205,31 +205,31 @@ private[spark] object GradientBoostedTrees extends
Logging {
case _ => data
}
- val numIterations = trees.length
- val evaluationArray = Array.fill(numIterations)(0.0)
- val localTreeWeights = treeWeights
-
- var predictionAndError = computeInitialPredictionAndError(
- remappedData, localTreeWeights(0), trees(0), loss)
-
- evaluationArray(0) = predictionAndError.values.mean()
-
val broadcastTrees = sc.broadcast(trees)
- (1 until numIterations).foreach { nTree =>
- predictionAndError =
remappedData.zip(predictionAndError).mapPartitions { iter =>
- val currentTree = broadcastTrees.value(nTree)
- val currentTreeWeight = localTreeWeights(nTree)
- iter.map { case (point, (pred, error)) =>
- val newPred = updatePrediction(point.features, pred,
currentTree, currentTreeWeight)
- val newError = loss.computeError(newPred, point.label)
- (newPred, newError)
- }
+ val localTreeWeights = treeWeights
+ val treesIndices = trees.indices
+
+ val dataCount = remappedData.count()
+ val evaluation = remappedData
+ .map { (point: LabeledPoint) =>
+ treesIndices
+ .map(idx => {
+ val prediction = broadcastTrees.value(idx)
+ .rootNode
+ .predictImpl(point.features)
+ .prediction
+ prediction * localTreeWeights(idx)
--- End diff --
It might be very slightly more consistent to try to also use
`updatePrediction` here with a 0 current prediction, just to make it pro forma
consistent with the rest.
---
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
