Wrong prediction for sentiment analysis

KNIME Analytics Platform
#1

Hello,

I just did training and testing for sentiment analysis and I want, if it’s possible to get to know which text was wrongly predicted. Because as of now, I can only can the document class and prediction in output of the Logistic Regression Predictor node but the ID I assigned manually to each text before processing was lost in the process.

Is it possible to get to know this information or is it just a lost cause ?

Thanks in advance for your help.

PS : the workflow I used was this one, I just changed the decision tree nodes by logistic regression nodes.

#2

Use Document Data Extractor at the end to get the text

1 Like
#3

Thank you for your hep @Daniel_Weikert

This node helps me to get the preprocessed document used in the predictor, but do you know of a way to get the starting document ?