Usage of ModelUpdater

jhahn

I want to make a naive bayes classifier that re-estimates its model iteratively.

Please refer to the image.



In my understanding on the "IteratingOperatorChain", the final ouput of the chain is passed to the input of the chain.
So, I have checked "keep_example_set" to pass the example set.
But, as you can see, an error is reported that "ModelUpdater (3): Missing input: ExampleSet"

I uploaded the project file.
http://ci.kaist.ac.kr/jhahn/test.xml

Please let me know what should I do.
Thanks in advance.

steffen

Hello

Ok ... I do not understand what you want to achieve.  From a RM technical point of view however:
- You have to mark "keep_example_set" in naivebayes
- select a performance measure in your classification performance operator

but this is not enough, because the model you have created before Iterating Operator Chain is not passed to the first child node of crossvalidation (select an operator and press F1 or refer to the tutorial.pdf to get more information about the operators). Use IORetriever and IOStorer to get the model into the right position.

Furthermore, there is an operator called "ModelUpdater" you could use instead of Modelgrouper.

regards,

Steffen

PS: I suggest to use the "#" option while creating a new message to post the xml-process file.