[SPARK-9647] MLlib + SparkR integration for 1.6

• Type: Umbrella
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Fix Version/s: 1.6.0
• Component/s: ML, SparkR
• Target Version/s: 1.6.0
• Assignee: Xiangrui Meng
• Reporter: Xiangrui Meng
• Created: 05/Aug/15 17:47
• Updated: 13/Nov/15 16:34
• Resolved: 13/Nov/15 16:34

Description

This is an umbrella JIRA for MLlib + SparkR integration for Spark 1.6, continuing the work from SPARK-6805. Some suggested features:

1. model coefficients for logistic regression
2. support feature interactions in RFormula
3. summary statistics -> implemented for linear regression
4. more error families -> 1.7

Issue Links

• depends upon: SPARK-6805 MLlib + SparkR integration for 1.5
• is related to: SPARK-9835 Iteratively reweighted least squares solver for GLMs
• is related to: SPARK-9698 Add feature interaction as a transformer
• is related to: SPARK-9834 Normal equation solver for ordinary least squares

Sub-Tasks

• LogisticRegression in R should provide model statistics: Resolved
• Support R feature interactions in RFormula: Resolved
• Provide R-like summary statistics for ordinary least squares via normal equation solver: Resolved
• Support transform string label for RFormula: Resolved
• R-like summary statistics with intercept for OLS via normal equation solver: Resolved
• Expose R-like summary statistics in SparkR::glm for linear regression: Resolved