[SPARK-23730] Save and expose "in bag" tracking for random forest model

• Type: Improvement
• Status: Resolved
• Priority: Minor
• Resolution: Incomplete
• Affects Version/s: 2.3.0
• Component/s: ML
• Labels: bulk-closed
• Assignee: Unassigned
• Reporter: Julian King
• Created: 18/Mar/18 10:07
• Updated: 08/Oct/19 05:41
• Resolved: 08/Oct/19 05:41

Description

In a random forest model, it is often useful to be able to keep track of which samples ended up in each of the bootstrap replications (and how many times this happened). For instance, in the R randomForest package this is accomplished through the option keep.inbag=TRUE

Similar functionality in Spark ML's random forest would be helpful

Issue Links

• Is contained by: SPARK-14046 RandomForest improvement umbrella