Using multiple cores with weka

[Robert Bixler]

Is it possible to use multiple cores to process multiple weka tasks
programmatically? For instance, I want to create a data set and then over a
certain number of iterations create a train and test set. I then want to
create a number of models using different classifieds and evaluate them. I
want to create a thread that does this process for each data set and run
each thread on a different core. Would I need to use weka server and set up
a server with the number of cores on my machine to do this?

[Mark Hall]

From: Robert Bixler <***@nd.edu<mailto:***@nd.edu>>
Reply-To: "Weka machine learning workbench list." <***@list.scms.waikato.ac.nz<mailto:***@list.scms.waikato.ac.nz>>
Date: Friday, 17 May 2013 7:38 AM
To: "Weka machine learning workbench list." <***@list.scms.waikato.ac.nz<mailto:***@list.scms.waikato.ac.nz>>
Subject: [Wekalist] Using multiple cores with weka

Is it possible to use multiple cores to process multiple weka tasks programmatically? For instance, I want to create a data set and then over a certain number of iterations create a train and test set. I then want to create a number of models using different classifieds and evaluate them. I want to create a thread that does this process for each data set and run each thread on a different core. Would I need to use weka server and set up a server with the number of cores on my machine to do this?

You could use Weka server or you could use the Experimenter's distributed experiment environment. In the later case you can start a RemoteEngine multiple times on the same machine (using different ports for RMI) and then in the client Experimenter's "Hosts" dialog specify your machine's hostname or IP address multiple times.

Cheers,
Mark.