ID3Numerical Bug??

adaman

Hey,
i´m not sure if this is a bug, but i´m searching for a tree model which is updateable and as it is writen in the JavaDOCS http://rapid-i.com/api/rapidminer-5.1/com/rapidminer/operator/learner/tree/ID3NumericalLearner.html
the Class ID3NumericalLearner comes from the AbstractTreeLearner and should be updatable. But my little example shows, that it does not.

The first code is for creation of the model, the second for the update. I use the latest Version 5.1.006. If i´m wrong i would appreciate if s.o. can tell me if there is any tree which is updatable or how to use W-BayesNet so that i can update.

Thank you very much for the help and greetings

<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<process version="5.1.006">
  <context>
    <input/>
    <output>
      <location>./ID3Model</location>
    </output>
    <macros/>
  </context>
  <operator activated="true" class="process" compatibility="5.1.006" expanded="true" name="Process">
    <process expanded="true" height="456" width="1006">
      <operator activated="true" class="generate_data" compatibility="5.1.006" expanded="true" height="60" name="Generate Data" width="90" x="112" y="75">
        <parameter key="target_function" value="polynomial classification"/>
      </operator>
      <operator activated="true" class="parallel:id3_numerical_parallel" compatibility="5.1.000" expanded="true" height="76" name="ID3 Numerical" width="90" x="380" y="75"/>
      <connect from_op="Generate Data" from_port="output" to_op="ID3 Numerical" to_port="training set"/>
      <connect from_op="ID3 Numerical" from_port="model" to_port="result 1"/>
      <portSpacing port="source_input 1" spacing="0"/>
      <portSpacing port="sink_result 1" spacing="0"/>
      <portSpacing port="sink_result 2" spacing="0"/>
    </process>
  </operator>
</process>

<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<process version="5.1.006">
  <context>
    <input>
      <location>./ID3Model</location>
    </input>
    <output/>
    <macros/>
  </context>
  <operator activated="true" class="process" compatibility="5.1.006" expanded="true" name="Process">
    <process expanded="true" height="474" width="1024">
      <operator activated="true" class="generate_data" compatibility="5.1.006" expanded="true" height="60" name="Generate Data" width="90" x="41" y="119">
        <parameter key="target_function" value="polynomial classification"/>
      </operator>
      <operator activated="true" class="update_model" compatibility="5.1.006" expanded="true" height="76" name="Update Model" width="90" x="246" y="75"/>
      <connect from_port="input 1" to_op="Update Model" to_port="model"/>
      <connect from_op="Generate Data" from_port="output" to_op="Update Model" to_port="example set"/>
      <connect from_op="Update Model" from_port="example set" to_port="result 1"/>
      <portSpacing port="source_input 1" spacing="0"/>
      <portSpacing port="source_input 2" spacing="0"/>
      <portSpacing port="sink_result 1" spacing="0"/>
      <portSpacing port="sink_result 2" spacing="0"/>
    </process>
  </operator>
</process>

Marco_Boeck

Hi,

you've posted the same code block twice..
Apart from that, where did you get the idea that the ID3NumericalLearner is updatable?
However there is an W-NaiveBayesUpdateable operator, which should be the one you're looking for.

Regards,
Marco

adaman

Ups my fault, i have corrected it.

I know that the NaiveBayes can be updated. But to say it with the words of my Prof. "You know it´s Naive" :-)
So i´m looking for any other Modeling Operator that is ready for update. Therefor i have taken  a look into the JAVADOCS and thought that the ID3 is also updatable as it has the same functions like fireUpdate as the NaiveBayes has.

So if i´m wrong, my fault. But where can i see or find out which Modelling operators are ready for update?

Greetings

Marco_Boeck

Hi,

the fireUpdate is just for event handling, nothing to do with updating the model
However I must admit that I don't know which models are updatable, so I'm afraid you'll have to play around and check the models yourself.
After a quick glance though it looks like no "standard" RapidMiner models are updatable, so try the WEKA extension models first.

Regards,
Marco

adaman

Thanks for the help,
with a little thinking i could have figure it out on my own, that  fireUpdate is for event handling :-)

Anyway even none (or most) of the weka models are updateable, so i had to change the hole process.

Thanks again