Options

CHAID relevance criterion

franaffranaf Member Posts: 13 Contributor II
edited November 2018 in Help
Hi, I'm interesting in using CHAID as operator in a process but when I start it, it gives an error message: "No relevance criterion defined". I couldn't find where where the problem is.
Thank you for your help.

Answers

  • Options
    landland RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 2,531 Unicorn
    Hi,
    which version of RapidMiner are you using? In the most current version I cannot reproduce this problem easily.
    If you are using version 5.0.003, please build a process where you use only data generators and that still reproduces the problem. If you could post it here, I would be able to find the bug if it still exists.

    Greetings,
      Sebastian
  • Options
    franaffranaf Member Posts: 13 Contributor II
    Hi, thank you for answering.
    I've installed 5.0.003 and now it's reporting another error. It says "numerical attributes not supported". I've used "Generate Up-selling Data" as automatic data generator and the problem persists. I've logged the process with logverbosity = "all". The results are here:
    Mar 4, 2010 11:58:06 AM com.rapidminer.Process prepareRun
    FINE: Initialising process setup.
    Mar 4, 2010 11:58:06 AM com.rapidminer.tools.WrapperLoggingHandler log
    INFO: No filename given for result file, using stdout for logging results!
    Mar 4, 2010 11:58:06 AM com.rapidminer.Process prepareRun
    FINE: Process initialised.
    Mar 4, 2010 11:58:06 AM com.rapidminer.Process loadInitialData
    INFO: Loading initial data.
    Mar 4, 2010 11:58:06 AM com.rapidminer.Process loadInitialData
    FINE: Input #1 not specified.
    Mar 4, 2010 11:58:06 AM com.rapidminer.Process run
    INFO: Process starts
    Mar 4, 2010 11:58:06 AM com.rapidminer.Process run
    FINE: Process:
      Process[1] (Process)
        subprocess 'Main Process'
          +- Generate Up-Selling Data[1] (Generate Up-Selling Data)
          +- CHAID[1] (CHAID)
    Mar 4, 2010 11:58:06 AM com.rapidminer.operator.Operator execute
    FINE: Starting application 1 of operator Process
    Mar 4, 2010 11:58:06 AM com.rapidminer.operator.Operator execute
    FINER: Process called 1st time with input:
    Mar 4, 2010 11:58:06 AM com.rapidminer.operator.execution.SimpleUnitExecutor execute
    FINE: Executing subprocess Process.Main Process. Execution order is: [Generate Up-Selling Data (Generate Up-Selling Data), CHAID (CHAID)]
    Mar 4, 2010 11:58:06 AM com.rapidminer.operator.Operator execute
    FINE: Starting application 1 of operator Generate Up-Selling Data
    Mar 4, 2010 11:58:06 AM com.rapidminer.operator.Operator execute
    FINER: Generate Up-Selling Data called 1st time with input:
    Mar 4, 2010 11:58:06 AM com.rapidminer.operator.Operator execute
    FINE: Completed application 1 of operator Generate Up-Selling Data
    Mar 4, 2010 11:58:06 AM com.rapidminer.operator.Operator execute
    FINER: Generate Up-Selling Data returned with output:
      outputSimpleExampleSet:
    100 examples,
    8 regular attributes,
    special attributes = {
        label = #8: label (nominal/single_value)/values=[product_1, product_2, product_3]
    }
    Mar 4, 2010 11:58:06 AM com.rapidminer.operator.Operator execute
    FINEST: Generate Up-Selling Data: execution time was 2 ms
    Mar 4, 2010 11:58:06 AM com.rapidminer.operator.Operator execute
    FINE: Starting application 1 of operator CHAID
    Mar 4, 2010 11:58:06 AM com.rapidminer.operator.Operator execute
    FINER: CHAID called 1st time with input:
      training setSimpleExampleSet:
    100 examples,
    8 regular attributes,
    special attributes = {
        label = #8: label (nominal/single_value)/values=[product_1, product_2, product_3]
    }
    Mar 4, 2010 11:58:06 AM com.rapidminer.operator.learner.tree.CHAIDLearner$1 getNominalBenefit
    INFO: Calc benefit for: SimpleExampleSet:
    100 examples,
    1 regular attributes,
    special attributes = {
        label = #8: label (nominal/single_value)/values=[product_1, product_2, product_3]
    }
    Mar 4, 2010 11:58:06 AM com.rapidminer.operator.learner.tree.CHAIDLearner$1 getNominalBenefit
    INFO: Weights are: AttributeWeights (containing weights for 1 attributes)
    Mar 4, 2010 11:58:06 AM com.rapidminer.operator.learner.tree.CHAIDLearner$1 getNominalBenefit
    INFO: W(name)=1.0
    Thank you very much
  • Options
    dragoljubdragoljub Member Posts: 241 Contributor II
    I get this same error. According to the operator formation it should be able to support numerical attributes.

    -Gagi
  • Options
    landland RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 2,531 Unicorn
    Hi,
    that's only an erroneous entry. We have removed that in the meanwhile. Sorry for the inconvenience.

    Greetings,
      Sebastian
  • Options
    franaffranaf Member Posts: 13 Contributor II
    Well, Sebastian.
    I've downloaded 005. Problem persists. Can't handle numerical values with CHAID.
  • Options
    landland RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 2,531 Unicorn
    Hi,
    I think you misunderstood my solution: Chaid simply does not support numerical attributes, so I removed the erroneous entry that it does! The description of it's behavior was changed, not the behavior itself.

    Greetings,
      Sebastian
  • Options
    franaffranaf Member Posts: 13 Contributor II
    Sebastian Land wrote:

    Hi,
    I think you misunderstood my solution: Chaid simply does not support numerical attributes, so I removed the erroneous entry that it does! The description of it's behavior was changed, not the behavior itself.

    Greetings,
      Sebastian
    OK. Other implementations of CHAID does.
    Thank you
Sign In or Register to comment.