[SPARK-8484] Add TrainValidationSplit to ml.tuning

• Type: New Feature
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Fix Version/s: 1.5.0
• Component/s: ML
• Target Version/s: 1.5.0
• Assignee: Martin Zapletal
• Reporter: Xiangrui Meng
• Shepherd: Xiangrui Meng
• Created: 19/Jun/15 17:28
• Updated: 12/Aug/15 22:10
• Resolved: 23/Jul/15 00:35

Description

Add TrainValidationSplit for hyper-parameter tuning. It randomly splits the input dataset into train and validation and use evaluation metric on the validation set to select the best model. It should be similar to CrossValidator, but simpler and less expensive.

Issue Links

• relates to: SPARK-8971 Support balanced class labels when splitting train/cross validation sets
• relates to: SPARK-8983 ML Tuning Cross-Validation Improvements
• relates to: SPARK-9910 User guide for train validation split
• links to: [Github] Pull Request #6996 (zapletal-martin)
• links to: [Github] Pull Request #7337 (zapletal-martin)