MOA in production environment

[Xavier Daull]

I am a frequent user of Weka and recently discovered MOA while searching for stream processing. It really provided me a far better understanding on my selected classifier over the time (mainly on using adaptive sliding window and concept drift detection).

But now I really wonder how to use my analysis' results from MOA with wrapped weka's RF and set up a real world environment with weka with concept drift detection and sliding windows. How is MOA used once in the production environment ? Do you have any command line or java code showing this kind of usage ?

Thank you

[Albert Bifet]

Yes, you can use MOA from your Java source code. Take a look at Tutorial 2:

http://sourceforge.net/projects/moa-datastream/files/documentation/Tutorial2.pdf/download

Cheers, Albert

> --
> You received this message because you are subscribed to the Google Groups
> "MOA users" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to moa-users+...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.

[Xavier Daull]

Thanks Albert. Now, I have been able to use several classifiers from Weka but I always get an error when trying to use RandomForest. I know it is not an updatable classifier but while using InterleavedTestThenTrain, I thought it could rebuild the model each time instead of updating it ? Or is the problem elsewhere ?

Regards,
Xavier

Failure reason: weka.classifiers.trees.RandomTree.setDoNotCheckCapabilities(Z)V
*** STACK TRACE ***java.lang.NoSuchMethodError: weka.classifiers.trees.RandomTree.setDoNotCheckCapabilities(Z)V
    at weka.classifiers.trees.RandomForest.buildClassifier(RandomForest.java:620)
    at moa.classifiers.meta.WEKAClassifier.buildClassifier(WEKAClassifier.java:152)
    at moa.classifiers.meta.WEKAClassifier.trainOnInstanceImpl(WEKAClassifier.java:109)
    at moa.classifiers.AbstractClassifier.trainOnInstance(AbstractClassifier.java:142)
    at moa.tasks.EvaluateInterleavedTestThenTrain.doMainTask(EvaluateInterleavedTestThenTrain.java:137)
    at moa.tasks.MainTask.doTaskImpl(MainTask.java:50)
    at moa.tasks.AbstractTask.doTask(AbstractTask.java:57)
    at moa.tasks.TaskThread.run(TaskThread.java:76)

[Albert Bifet]

It seems that the system can not found a method:

[Albert Bifet]

It seems that the system can not found a method:

java.lang.NoSuchMethodError

Which version of weka are you using?

Cheers, Albert