[
https://issues.apache.org/jira/browse/FLINK-2157?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14618303#comment-14618303
]
ASF GitHub Bot commented on FLINK-2157:
---------------------------------------
Github user tillrohrmann commented on a diff in the pull request:
https://github.com/apache/flink/pull/871#discussion_r34131445
--- Diff:
flink-staging/flink-ml/src/main/scala/org/apache/flink/ml/package.scala ---
@@ -70,6 +70,15 @@ package object ml {
dataSet.map(new BroadcastSingleElementMapperWithIteration[T, B,
O](dataSet.clean(fun)))
.withBroadcastSet(broadcastVariable, "broadcastVariable")
}
+
+ /** Calculates the mean value of a DataSet[T <\: Numeric[T]]
+ *
+ * @return A DataSet[Double] with the mean value as its only element
+ */
+ def mean()(implicit num: Numeric[T], ttit: TypeInformation[(T, Int)]):
DataSet[Double] =
--- End diff --
I think this method should be placed in a `NumericRichDataSet[T: Numeric]`
class. Otherwise we provide a `mean` method for `DataSets` whose element type
does not support this operation. This might be misleading IMO.
> Create evaluation framework for ML library
> ------------------------------------------
>
> Key: FLINK-2157
> URL: https://issues.apache.org/jira/browse/FLINK-2157
> Project: Flink
> Issue Type: New Feature
> Components: Machine Learning Library
> Reporter: Till Rohrmann
> Assignee: Theodore Vasiloudis
> Labels: ML
> Fix For: 0.10
>
>
> Currently, FlinkML lacks means to evaluate the performance of trained models.
> It would be great to add some {{Evaluators}} which can calculate some score
> based on the information about true and predicted labels. This could also be
> used for the cross validation to choose the right hyper parameters.
> Possible scores could be F score [1], zero-one-loss score, etc.
> Resources
> [1] [http://en.wikipedia.org/wiki/F1_score]
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
