[
https://issues.apache.org/jira/browse/SPARK-1791?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Xiangrui Meng resolved SPARK-1791.
----------------------------------
Resolution: Fixed
Fix Version/s: 1.0.0
PR: https://github.com/apache/spark/pull/725
> SVM implementation does not use threshold parameter
> ---------------------------------------------------
>
> Key: SPARK-1791
> URL: https://issues.apache.org/jira/browse/SPARK-1791
> Project: Spark
> Issue Type: New Feature
> Components: MLlib
> Reporter: Andrew Tulloch
> Fix For: 1.0.0
>
> Original Estimate: 10m
> Remaining Estimate: 10m
>
> The key error is in SVM.scala, in `predictPoint`
> ```
> threshold match {
> case Some(t) => if (margin < 0.0) 0.0 else 1.0
> case None => margin
> }
> ```
--
This message was sent by Atlassian JIRA
(v6.2#6252)
