You can always use the ml libs in R, but you have to integrate them in
sparkr (= make all the logic to run in parallel etc). However, for your use
case it may make more sense to write the wrapper R mllib yourself, if the
project cannot provide it in time. It is not that difficult to call java or
scala classes from R.

Le ven. 4 sept. 2015 à 14:50, Jonathan Hodges <[email protected]> a écrit :

> Hi Spark Experts,
>
> We are trying to streamline the development lifecycle of our data
> scientists taking algorithms from the lab into production.  Currently the
> tool of choice for our data scientists is R.  Historically our engineers
> have had to manually convert the R based algorithms to Java or Scala to run
> in production on Hadoop or Spark clusters.
>
> We are curious if we can do better by leveraging SparkR or MLlib by data
> scientists to minimize the manual translation to move algorithms into
> production.  Ideally it would be great to use SparkR as the data scientists
> are much more familiar with R than MLlib.  Can SparkR run in production or
> are there some downsides to this approach?
>
> I noticed the following JIRAs for MLlib / SparkR integration.
>
> https://issues.apache.org/jira/browse/SPARK-6805
> https://issues.apache.org/jira/browse/SPARK-9647
>
> Beyond the lack of full MLlib features supported in SparkR, the main
> question is if it is as stable and fault tolerant as using MLlib directly.
>
> Thanks in advance for any guidance you can provide.
>
> Jonathan
>
>

Reply via email to