CODEWITHSUNDEEP
javartomcatpmml
SoakingHummer
SoakingHummer

Reputation: 572

How to set up a system that reads PMML files which I create using R?

For the past few days I'm stuck at trying to pass a PMML file with .xml extension that I created using R and pmml package. After creating a iris.xml PMML file I have no where to go, actually I don't know what to look for. And then I saw Alex Guazzelli's video here, where he tells us to use ADAPA from Zementis to run the PMML files that are created from R.

What I took away from the video is,

  1. I want to do something similar as shown, like uploading my PMML file at some place and input my test files frequently to get the corresponding predicted values.
  2. ADAPA looks great, but looks like it also costs some money. So I would like to look for an alternative, which might not be as user friendly as ADAPA.
  3. Then I found this page where the author says,

So the first step was to get the engine running. I experienced some small problems but Scott fixed them very fast and offered a new war file ready for deployment inside tomcat. It can be obtained under http://code.google.com/p/openscoring/. If you still encounter some problems you could also try my war file. It was built with java version “1.6.0_26” and tested under tomcat 7. After installing the war file using the management console, you could test it by sending a POST request in XML like the following to ‘http://localhost:8080/OpenScoring/Scoring’

Then I found jpmml. The problem here is the package names are very confusing, I don't know which one to install. In fact I don't know how to install each of them. I have basic knowledge of Java, apparently it isn't helpful in the current scenario.

Desperately looking forward for any kind of help.

Upvotes: 1

Views: 1000

Answers (1)

user1808924
user1808924

Reputation: 4926

How do the data flows in your application look like? Zementis' ADAPA is a standalone web service, which lets you do scoring in a request/response fashion. This works for web applications, but is clearly sub-optimal if you're looking to score larger amounts of data.

The JPMML project provides "PMML runtimes" for most popular big data frameworks. For example, if your application is tied to Apache Spark, then you would go for the JPMML-Spark project, if tied to Apache Hive/Pig, then for the JPMML-Pig project, etc. For a simple REST web service you should check out the Openscoring project, which is functionally identical to Zementis' ADAPA.

All JPMML projects can be built from their source checkout using Apache Maven by issuing the following command:

mvn clean install

Some JPMML projects also have their binaries distributed via the Maven Central repository. If you're looking for the Openscoring WAR file, then you need to search for "openscoring-webapp" and on the results page click the "download WAR" link (at the time of writing this, you would get openscoring-webapp-1.2.12.war). This is a standard WAR file that can be dropped to the auto deploy folder of your favourite Java web server (Tomcat?) as it is. The REST API of the web service is fully documented in the README file of the Openscoring project.

Upvotes: 1

Related Questions