Hitachi Vantara Pentaho Community Forums
Results 1 to 2 of 2

Thread: Parameters for J48 Tree

  1. #1
    Join Date
    Apr 2016

    Default Parameters for J48 Tree


    I'm using the J48 Tree for classification. It support lots of parameters.

    Which parameter ranges should I search for? I'm especially unsure about the following settings:

    runing confidence
    minimum number of instances

    Second, how can I implement Bagging with a J48 classifier in WEKA?

    EDIT: I have just seen that some methods like buildClassifier() throws an Exception of type Exception. Why in gods name is this done? Throwing these unspecific exceptions is not so good programming style.
    Last edited by BlackHawk; 04-24-2016 at 05:35 AM.

  2. #2
    Join Date
    Aug 2006


    Default settings work well for most datasets. I wouldn't bother fiddling with the pruning confidence, and the minimum number of instances is only worth adjusting if you are going to turn off post pruning. Post pruning is more powerful than pre-pruning (early stopping), which is what the minimum number of instances controls. If you really want to find the optimal settings for J48 on a particular dataset then you can use a meta-learner like GridSearch or CVParameterSelection.

    Bagging in Weka can be applied to any base classifier. The default base classifier is REPTree, but you can change it to J48 in the Explorer or by using the -W option from the command line. Bagging decision trees works best when pruning is turned off and the option for Laplace smoothing of probability estimates is turned on.


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.