[SPARK-5016] GaussianMixtureEM should distribute matrix inverse for large numFeatures, k

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.2.0
• Fix Version/s: 1.5.0
• Component/s: MLlib
• Labels: clustering
• Target Version/s: 1.5.0
• Assignee: Feynman Liang
• Reporter: Joseph K. Bradley
• Shepherd: Xiangrui Meng
• Created: 30/Dec/14 20:39
• Updated: 08/Jul/15 23:32
• Resolved: 08/Jul/15 23:32

Description

If numFeatures or k are large, GMM EM should distribute the matrix inverse computation for Gaussian initialization.

Issue Links

• is required by: SPARK-7206 Gaussian Mixture Model (GMM) improvements
• links to: [Github] Pull Request #4654 (MechCoder)
• links to: [Github] Pull Request #7166 (feynmanliang)