Hello KNIME community.
Hoping for some advice, I’m trying to predict TargetBuy (Organics Purchase Indicator)
- I’m wondering why my accuracy is 100% for all models?
- Also, must all data be converted to a string to process?
- Any way to be more efficient in this? i.e using other nodes
Data Preparation_Trial.knwf (75.5 KB)
Organics.xlsx (1.3 MB)
Hi @Matthias_CTS and welcome to the forum.
In this case it looks like this happens because you have included the field Organics Purchase Count as a feature. If you remove it, your results look a lot more reasonable.
EDIT: By the way, one way to intuit this is to look at the view of the decision tree. There you can see that the model is taking only a single predictor into account to arrive at 100% accuracy.
This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.