Hitachi Vantara Pentaho Community Forums
Results 1 to 2 of 2

Thread: How to run Weka faster?

  1. #1
    Join Date
    Jan 2016
    Posts
    4

    Default How to run Weka faster?

    Hello all,
    I am running Weka (3.7.12) on MacBook Pro with 16 GB Ram. I am running attribute selection with Wrapper method using Random Forest. My dataset has 19 features only and 100580 instances. I changed info.plist and increased the memory to be used as:
    <array>
    <string>-Xmx8192M</string> (I can put more RAM if that helps)
    </array>
    But the wrapper method takes forever to run. Is there any way to make it run faster?

    Thanks for any help.

    Regards,
    Raj

  2. #2
    Join Date
    Aug 2006
    Posts
    1,741

    Default

    Try limiting the depth of the trees that Random forest is generating. The default is unlimited depth - actually limited by a min number of instances at the leaves, but in your case (assuming numeric attributes) this could generate really big trees. You could also limit the number of trees learned. Note that wrapper subset evaluation is fairly slow, due to the fact that a repeated 5-fold cross validation is used internally on the training data in order to evaluate the merit of individual subsets.

    Also, if you are using Weka 3.7 then RandomForest has an option to run in multiple threads.

    Cheers,
    Mark.
    Last edited by Mark; 02-11-2016 at 04:15 AM.

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
Privacy Policy | Legal Notices | Safe Harbor Privacy Policy

Copyright © 2005 - 2019 Hitachi Vantara Corporation. All Rights Reserved.