[HIVEMALL-243] Fix nominal variable handling in DecisionTree and RegressionTree

• Type: Task
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 0.5.2
• Fix Version/s: 0.6.0
• Assignee: Makoto Yui
• Reporter: Makoto Yui
• Created: 11/Mar/19 09:00
• Updated: 13/Mar/19 07:56
• Resolved: 13/Mar/19 07:56

Description

For NOMINAL variable, the maximum attribute index 'm' is used for computing splits.

This cause performance issues for sparse nominal variables. So, revise this handling for a better performance.

https://github.com/apache/incubator-hivemall/blob/master/core/src/main/java/hivemall/smile/classification/DecisionTree.java#L703

Issue Links

• links to: GitHub Pull Request #185