[
https://issues.apache.org/jira/browse/SPARK-3181?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16181724#comment-16181724
]
Joseph K. Bradley commented on SPARK-3181:
------------------------------------------
Re: [~sethah]'s comment about separating Huber Estimator from regular
LinearRegression in the PR:
This was my initial reaction too, but I can see both sides:
* Technically, robust regression using the Huber loss is Linear Regression. As
far as I know (and as far as I can tell from Wikipedia), "Linear Regression"
refers to the predictive model, not to the loss model. Using Huber loss
instead of squared error does not change the predictive model.
* Users should get what they expect when they use an Estimator. The average
user would not expect Linear Regression to do fancy Huber loss regression.
That said, the default behavior *would* be least squares regression, so it's
not a real problem.
I don't have strong feelings here, but I'm fine with them being combined.
Thinking about the past, it was definitely overkill to separate
LinearRegression, RidgeRegression, and Lasso in the old RDD-based API.
> Add Robust Regression Algorithm with Huber Estimator
> ----------------------------------------------------
>
> Key: SPARK-3181
> URL: https://issues.apache.org/jira/browse/SPARK-3181
> Project: Spark
> Issue Type: New Feature
> Components: ML
> Affects Versions: 2.2.0
> Reporter: Fan Jiang
> Assignee: Yanbo Liang
> Labels: features
> Original Estimate: 0h
> Remaining Estimate: 0h
>
> Linear least square estimates assume the error has normal distribution and
> can behave badly when the errors are heavy-tailed. In practical we get
> various types of data. We need to include Robust Regression to employ a
> fitting criterion that is not as vulnerable as least square.
> In 1973, Huber introduced M-estimation for regression which stands for
> "maximum likelihood type". The method is resistant to outliers in the
> response variable and has been widely used.
> The new feature for MLlib will contain 3 new files
> /main/scala/org/apache/spark/mllib/regression/RobustRegression.scala
> /test/scala/org/apache/spark/mllib/regression/RobustRegressionSuite.scala
> /main/scala/org/apache/spark/examples/mllib/HuberRobustRegression.scala
> and one new class HuberRobustGradient in
> /main/scala/org/apache/spark/mllib/optimization/Gradient.scala
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
