Issue with replicating (multiply operator) test data

amitd

I am analyzing the churn dataset (WA_Fn UseC_ Telco Customer Churn.csv) from theIBM sample datasets website. Attached is the XML process and the error encountered when I try to replicate the test data port from the CV operator. I am using 3 replications of the test data, one for creating a lift chart, another one for creating ROCs and yet another one for simply as output (to see the predictions and the confidence values). However, the output of the predictions and confidence values are not coming through. Testing with breakpoints show that the test data before and after the multiply operator has those values properly generated, but the final output does not display it. Any ideas would be of much help.

<反对text>






<运营商激活= " true "类=“检索”兼容ibility="9.0.002" expanded="true" height="68" name="Retrieve WA_Fn-UseC_-Telco-Customer-Churn" width="90" x="112" y="85">
<参数键= " repository_entry " value = " . . /数据/ WA_Fn-UseC_-Telco-Customer-Churn"/>





























<反对nect from_port="training set" to_op="Decision Tree" to_port="training set"/>
<反对nect from_op="Decision Tree" from_port="model" to_port="model"/>
<反对nect from_op="Decision Tree" from_port="weights" to_port="through 1"/>












<反对nect from_port="model" to_op="Apply Model" to_port="model"/>
<反对nect from_port="test set" to_op="Apply Model" to_port="unlabelled data"/>
<反对nect from_op="Apply Model" from_port="labelled data" to_op="Performance" to_port="labelled data"/>
<反对nect from_op="Performance" from_port="performance" to_port="performance 1"/>
<反对nect from_op="Performance" from_port="example set" to_port="test set results"/>





























<反对nect from_port="train 1" to_op="Decision Tree (2)" to_port="training set"/>
<反对nect from_port="train 2" to_op="Decision Tree (3)" to_port="training set"/>
<反对nect from_port="train 3" to_op="Decision Tree (4)" to_port="training set"/>
<反对nect from_op="Decision Tree (2)" from_port="model" to_port="model 1"/>
<反对nect from_op="Decision Tree (3)" from_port="model" to_port="model 2"/>
<反对nect from_op="Decision Tree (4)" from_port="model" to_port="model 3"/>










<反对nect from_op="Retrieve WA_Fn-UseC_-Telco-Customer-Churn" from_port="output" to_op="Nominal to Binominal" to_port="example set input"/>
<反对nect from_op="Nominal to Binominal" from_port="example set output" to_op="Numerical to Binominal" to_port="example set input"/>
<反对nect from_op="Numerical to Binominal" from_port="example set output" to_op="Set Role" to_port="example set input"/>
<反对nect from_op="Set Role" from_port="example set output" to_op="Multiply" to_port="input"/>
<反对nect from_op="Multiply" from_port="output 1" to_op="Cross Validation" to_port="example set"/>
<反对nect from_op="Cross Validation" from_port="model" to_op="Model Replicate" to_port="input"/>
<反对nect from_op="Cross Validation" from_port="test result set" to_op="Test data replicate" to_port="input"/>
<反对nect from_op="Cross Validation" from_port="performance 1" to_port="result 1"/>
<反对nect from_op="Model Replicate" from_port="output 1" to_port="result 2"/>
<反对nect from_op="Model Replicate" from_port="output 2" to_op="Lift Chart (Simple)" to_port="model"/>
<反对nect from_op="Test data replicate" from_port="output 1" to_port="result 3"/>
<反对nect from_op="Test data replicate" from_port="output 2" to_op="Lift Chart (Simple)" to_port="test data"/>
<反对nect from_op="Test data replicate" from_port="output 3" to_op="Compare ROCs" to_port="example set"/>
<反对nect from_op="Lift Chart (Simple)" from_port="lift chart" to_port="result 5"/>
<反对nect from_op="Compare ROCs" from_port="rocComparison" to_port="result 4"/>

David_A

Hi,

the solution is to put aMaterialize Dataoperator before theCompare ROCs operator.

It forces RapidMiner to create a clean copy of the example set, so the other version aren't overridden.

Normally this isn't required to do explicitly, but in some rare cases like this it is required.

Best,

David

amitd

Thanks, the general idea worked. I ended up putting the Materialize Data operator simply before the final output.

Here is the working solution for anyone else.

<反对text>






<运营商激活= " true "类=“检索”兼容ibility="9.0.002" expanded="true" height="68" name="Retrieve WA_Fn-UseC_-Telco-Customer-Churn" width="90" x="45" y="85">
<参数键= " repository_entry " value = " . . /数据/ WA_Fn-UseC_-Telco-Customer-Churn"/>




























<反对nect from_port="training set" to_op="Decision Tree" to_port="training set"/>
<反对nect from_op="Decision Tree" from_port="model" to_port="model"/>
<反对nect from_op="Decision Tree" from_port="weights" to_port="through 1"/>












<反对nect from_port="model" to_op="Apply Model" to_port="model"/>
<反对nect from_port="test set" to_op="Apply Model" to_port="unlabelled data"/>
<反对nect from_op="Apply Model" from_port="labelled data" to_op="Performance" to_port="labelled data"/>
<反对nect from_op="Performance" from_port="performance" to_port="performance 1"/>
<反对nect from_op="Performance" from_port="example set" to_port="test set results"/>






























<反对nect from_port="train 1" to_op="Decision Tree (2)" to_port="training set"/>
<反对nect from_port="train 2" to_op="Decision Tree (3)" to_port="training set"/>
<反对nect from_port="train 3" to_op="Decision Tree (4)" to_port="training set"/>
<反对nect from_op="Decision Tree (2)" from_port="model" to_port="model 1"/>
<反对nect from_op="Decision Tree (3)" from_port="model" to_port="model 2"/>
<反对nect from_op="Decision Tree (4)" from_port="model" to_port="model 3"/>










<反对nect from_op="Retrieve WA_Fn-UseC_-Telco-Customer-Churn" from_port="output" to_op="Nominal to Binominal" to_port="example set input"/>
<反对nect from_op="Nominal to Binominal" from_port="example set output" to_op="Numerical to Binominal" to_port="example set input"/>
<反对nect from_op="Numerical to Binominal" from_port="example set output" to_op="Set Role" to_port="example set input"/>
<反对nect from_op="Set Role" from_port="example set output" to_op="Cross Validation" to_port="example set"/>
<反对nect from_op="Cross Validation" from_port="model" to_op="Model Replicate" to_port="input"/>
<反对nect from_op="Cross Validation" from_port="test result set" to_op="Multiply" to_port="input"/>
<反对nect from_op="Model Replicate" from_port="output 1" to_port="result 1"/>
<反对nect from_op="Model Replicate" from_port="output 2" to_op="Lift Chart (Simple)" to_port="model"/>
<反对nect from_op="Multiply" from_port="output 1" to_op="Lift Chart (Simple)" to_port="test data"/>
<反对nect from_op="Multiply" from_port="output 2" to_op="Compare ROCs" to_port="example set"/>
<反对nect from_op="Multiply" from_port="output 3" to_op="Materialize Data" to_port="example set input"/>
<反对nect from_op="Materialize Data" from_port="example set output" to_port="result 4"/>
<反对nect from_op="Lift Chart (Simple)" from_port="lift chart" to_port="result 2"/>
<反对nect from_op="Compare ROCs" from_port="rocComparison" to_port="result 3"/>

sgenzer

this is very nice,@amitdeokar. Can I put this on the Community Repository?

Scott

amitd

Certainly, happy to share.

sgenzer

awesome. Thanks very much@amitdeokar. It's now in the Community Repo:

电信客户流失的用例

Scott