zhengruifeng commented on a change in pull request #27070:
[SPARK-9612][ML][FOLLOWUP] fix GBT support weights if subsamplingRate<1
URL: https://github.com/apache/spark/pull/27070#discussion_r362688433
##########
File path:
mllib/src/test/scala/org/apache/spark/ml/tree/impl/BaggedPointSuite.scala
##########
@@ -91,8 +91,7 @@ class BaggedPointSuite extends SparkFunSuite with
MLlibTestSparkContext {
baggedRDD.map(_.subsampleCounts.map(_.toDouble)).collect()
EnsembleTestHelper.testRandomArrays(subsampleCounts, numSubsamples,
expectedMean,
expectedStddev, epsilon = 0.01)
- // should ignore weight function for now
- assert(baggedRDD.collect().forall(_.sampleWeight === 1.0))
+ assert(baggedRDD.collect().forall(_.sampleWeight === 2.0))
Review comment:
because this testsuite meet conditions: withReplacement=false,
numSubsamples!=1,
it will call the modified `convertToBaggedRDDSamplingWithoutReplacement`,
and the `extractSampleWeight` here is `(_: LabeledPoint) => 2.0`, so output
baggedPoints will have sampleWeight==2.0
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
