[
https://issues.apache.org/jira/browse/SPARK-20810?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16017273#comment-16017273
]
Sean Owen commented on SPARK-20810:
-----------------------------------
Are you pretty sure both are converged?
You set the same params but do they have the same meaning in both
implementations?
I wonder if you can double-check the loss that both are computing to see if
they even agree about how good a solution the other has found.
I doubt the discontinuity of the hinge loss matters as it only affects the
gradient when the loss is exactly 0, and defining the derivative as 0 or 1 is
valid and doesn't matter much, or shouldn't.
> ML LinearSVC vs MLlib SVMWithSGD output different solution
> ----------------------------------------------------------
>
> Key: SPARK-20810
> URL: https://issues.apache.org/jira/browse/SPARK-20810
> Project: Spark
> Issue Type: Question
> Components: ML, MLlib
> Affects Versions: 2.2.0
> Reporter: Yanbo Liang
>
> Fitting with SVM classification model on the same dataset, ML {{LinearSVC}}
> produces different solution compared with MLlib {{SVMWithSGD}}. I understand
> they use different optimization solver (OWLQN vs SGD), does it make sense to
> converge to different solution? Since we use {{sklearn.svm.LinearSVC}} and R
> e1071 SVM as the reference in {{LinearSVCSuite}}, it seems like
> {{SVMWithSGD}} produce wrong solution. Does it also like this?
> AFAIK, both of them use {{hinge loss}} which is convex but not differentiable
> function. Since the derivative of the hinge loss at certain place is
> non-deterministic, should we switch to use {{squared hinge loss}} which is
> the default loss function of {{sklearn.svm.LinearSVC}} and more robust than
> {{hinge loss}}?
> This issue is very easy to reproduce, you can paste the following code
> snippet to {{LinearSVCSuite}} and then click run in Intellij IDE.
> {code}
> test("LinearSVC vs SVMWithSGD") {
> import org.apache.spark.mllib.linalg.{Vectors => OldVectors}
> import org.apache.spark.mllib.regression.{LabeledPoint => OldLabeledPoint}
> val trainer1 = new LinearSVC()
> .setRegParam(0.00002)
> .setMaxIter(200)
> .setTol(1e-4)
> val model1 = trainer1.fit(binaryDataset)
> println(model1.coefficients)
> println(model1.intercept)
> val oldData = binaryDataset.rdd.map { case Row(label: Double, features:
> Vector) =>
> OldLabeledPoint(label, OldVectors.fromML(features))
> }
> val trainer2 = new SVMWithSGD().setIntercept(true)
>
> trainer2.optimizer.setRegParam(0.00002).setNumIterations(200).setConvergenceTol(1e-4)
> val model2 = trainer2.run(oldData)
> println(model2.weights)
> println(model2.intercept)
> }
> {code}
> The output is:
> {code}
> [7.24661385022775,14.774484832179743,22.00945617480461,29.558498069476084]
> 7.373454363024084
> [0.58166680313823,1.1938960150473041,1.7940106824589588,2.4884300611292165]
> 0.667790514894194
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
