[
https://issues.apache.org/jira/browse/SPARK-5981?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14338936#comment-14338936
]
Joseph K. Bradley commented on SPARK-5981:
------------------------------------------
You can, but I'm going to split it into sub-tasks since it will require storing
the model within Python (simple for some models like Naive Bayes but complex
for others like trees). We can do one sub-task at a time.
Also, we'll have to see if there is a performance hit. I've been wondering
which is faster:
* store the model in Scala and do batch prediction on an RDD
* store the model in Python and do prediction on an RDD using map +
single-instance predict within the map
> pyspark ML models should support predict/transform on vector within map
> -----------------------------------------------------------------------
>
> Key: SPARK-5981
> URL: https://issues.apache.org/jira/browse/SPARK-5981
> Project: Spark
> Issue Type: Improvement
> Components: MLlib, PySpark
> Affects Versions: 1.3.0
> Reporter: Joseph K. Bradley
>
> Currently, most Python models only have limited support for single-vector
> prediction.
> E.g., one can call {code}model.predict(myFeatureVector){code} for a single
> instance, but that fails within a map for Python ML models and transformers
> which use JavaModelWrapper:
> {code}
> data.map(lambda features: model.predict(features))
> {code}
> This fails because JavaModelWrapper.call uses the SparkContext (within the
> transformation). (It works for linear models, which do prediction within
> Python.)
> Supporting prediction within a map would require storing the model and doing
> prediction/transformation within Python.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
