[SPARK-8538] LinearRegressionResults class for storing LR results on data

• Type: New Feature
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.5.0
• Component/s: ML
• Target Version/s: 1.5.0
• Assignee: Feynman Liang
• Reporter: Joseph K. Bradley
• Shepherd: Joseph K. Bradley
• Created: 22/Jun/15 21:20
• Updated: 09/Jul/15 23:21
• Resolved: 09/Jul/15 23:21

Description

This is an initial task for SPARK-7674. Please see the design doc linked from SPARK-7674 for details.

LinearRegressionResults will store results for a (model, dataset) pair, provided via lazy evaluation. It can take advantage of existing code such as RegressionMetrics.

Eventually, we may need to introduce an internal abstraction for RegressionResults, but that will not be necessary for this initial tasks. However, the assignee should make sure to coordinate with others working on SPARK-7674.

Note: This can add a limited number of metrics. We can always add more in the future.

Issue Links

• is required by: SPARK-7674 R-like stats for ML models
• is required by: SPARK-8539 LinearRegressionSummary class for storing LR training stats
• links to: [Github] Pull Request #7099 (feynmanliang)