[R] Network Alternative to rJava/JRI?

Jake Stone jake at jakestone.net
Wed Apr 26 15:55:26 CEST 2017

I am a java programmer, quite new to R.

I am familiar with rJava/JRI, but would prefer a distributed networked
architecture for my systems.


   - Java runs an NLP analysis of a text. It then sends an array (vector)
   of variables to R
   - R predicts the class of the input vector based on a pre-analyzed
   discriminant analysis of a full dataset in R.
   - R returns the class prediction to java.

R is an online webserver (perhaps shiny?) that "waits" for a request sent
via http. (This is similar to a jave servlet on a webserver)
Java program sends an http request with the array of variables
R responds with a response of the class prediction.

Is there an extant architecture that supports this. I expect shiny can do
this, but I am hesitant to spend two or three days learning shiny unless I
am confident the effort will pay off.



	[[alternative HTML version deleted]]

More information about the R-help mailing list