[
https://issues.apache.org/jira/browse/SPARK-11668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15001745#comment-15001745
]
Shubhanshu Mishra commented on SPARK-11668:
-------------------------------------------
Thanks just checked the package. The coefficient level summary stats are
presented. However, the model level summary stats are missing like AIC, BIC,
AUC, Precision, Recall, F-Measure, R-square etc. Probably the lrm package in R
or statsmodels package in python can be an inspiration for adding these summary
stats.
http://statsmodels.sourceforge.net/devel/examples/notebooks/generated/formulas.html
Maybe, even Weka model summary stats can also be used as inspiration for adding
functionality.
> R style summary stats in GLM package SparkR
> -------------------------------------------
>
> Key: SPARK-11668
> URL: https://issues.apache.org/jira/browse/SPARK-11668
> Project: Spark
> Issue Type: Improvement
> Components: SparkR
> Affects Versions: 1.5.0, 1.5.1
> Environment: LINUX
> WINDOWS
> MAC
> Reporter: Shubhanshu Mishra
> Labels: GLM, sparkr
> Fix For: 1.5.1
>
>
> In the current GLM module in R the `summary(model)` function call will only
> return the values of the coefficients however in the actual R GLM module, the
> function also returns the std. err, z score, p-value and confidence intervals
> for the coefficients as well as some model based statistics like R-squared
> values, AIC, BIC etc.
> Another inspiration for adding these metrics can be using the format of
> python statsmodels package described here:
> http://statsmodels.sourceforge.net/devel/examples/notebooks/generated/formulas.html
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
