Using model from pmml file to predict inside python script node

Hi everyone, I’ve been trying to use pypmml inside my script node to load a pmml file from an SVM trainer, but when I try to classify something it returns only unfavourable predictions (model accuracy is 75% normally). I was wondering if someone has had this issue before and/or if I am either misusing pypmml or whether I have to do something in Knime other than simply saving the pmml file and opening it later.

For anyone who happens to stumble upon this: I realized the issue was very simply that I was using a model that requires data to be normalized but I was asking it to classify a dataset that was not normalized!

2 Likes

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.