Decision Tree

Jason008

I use Decision Tree to  prediction  Ex use 102-105 (train) year to prediction 106  (test )  ,but 102-105 have 20 attribute 106 have ten , rapidminer  tell me attribute do not match , input exampleSet not match the training exampleSet .

i need to create 10 attribute on 106 but it is vacancy or 0 , then results is bad

ceaperez

Hi @Jason008, 
welcome to Rapidminer. 
You can't train your model using some attributes or predictos and then test it without that attributes. The trainned model will use that attributes as a part of the model and need the same attributes to test or apply the model. 
1- Evaluate the dataset and the extra 10 attributes and their relevance in the context of the problem. 
2- As a possible path you can review your dataset and try to complete your dataset 
3- You can evaluate a possible model using the 10 attributes that your dataset share.

Best, 

Cesar

Jason008

i see thx for help