How can I compare my prediction to the original label with the Validation Operator?

GLopes

Hello,

I'm doing a text mining project where I'm classifying text with a given label. I've split my data 70/30 with the Split Validation Operator. 

However, as this video "Finding the right model" on RapidMiner Youtube Page shows, I'm unable to see ofr which documents did my model fail. I can only see the confusion matrix... 

How can I see for each document which ones the model was wrong, as the Cross Validation operator does?

Best regards

varunm1

Hello @GLopes

Did you connect the "exa" of performance operator to the output "res"?

This will give you example set with original and predicted labels. You can analyze which are correct and wrong.

If you need more help just share you XML code by copying from (View --> Show Panel --> XML) and pasting here.

Hope this helps.

Varun

GLopes

Hello! Thank you so much for helping! I'm completely lost and I've seek the whole Internet for help on RapidMiner. 

I have the Performance Operator inside the Split validation, so it will only let me connect the "per" and the "exa" to "ave"... 

Here's the XML:  codeshare.io/UbKVU 
(too big to post her e:\ )

Telcontar120

Just output your model via the model port and then use another Apply Model operator to score the data (again) and then you will be able to do the comparison. 

GLopes

so, Telcontar120, inside the Validation, should I not use an Apply and a Performance? I'm confused..

Sorry it's my first RapidMiner process after the tutorials

sgenzer

hi @GLopes you may find the new Text Mining training resources on the RapidMiner Academy more useful and relevant to your project:

https://academy.rapidminer.com/courses/text-and-web-mining-with-rapidminer

Scott