Github user jkbradley commented on a diff in the pull request:
https://github.com/apache/spark/pull/3636#discussion_r22184450
--- Diff:
mllib/src/main/scala/org/apache/spark/mllib/optimization/GradientDescent.scala
---
@@ -77,6 +80,17 @@ class GradientDescent private[mllib] (private var
gradient: Gradient, private va
}
/**
+ * Set the convergence tolerance. Default 0.001
+ * convergenceTol is a condition which decides iteration termination.
+ * If the difference between last loss and last before loss is less than
convergenceTol
--- End diff --
The "solution vector" is ```weights```. (I agree with @mengxr on this.)
I'd vote for relative convergence tolerance such as:
sum_i (oldWeights[i] - newWeights[i])^2 / numWeights
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
