When updating a model created with a Weka learner, such as W-NaiveBayesMultinomialUpdateable, only every second example is used. This is due to the following lines in the method updateClassifier in WekaClassifier.java (which is indeed a Rapidminer operator, responsible for calling the Weka learners from Rapidminer):
for (int i = 0; i < instances.numInstances(); i++) {
Instance instance = instances.instance(i++);
classifier.updateClassifier(instance);
}
You are incrementing the variable i two times. You should fix that in the next version of Rapidminer.
Kind regards,
Daniel
thank you for this hint. I have corrected it and as soon as we merge our local repository with the sf-repository it can be checked out using the anonymous access.
Greetings,
Sebastian