Time Series Forecasting for Data with Compound Key

csl303

Hello Everyone,

 

I'm fairly new to Rapidminer and I would really appreciate your help handling a dataset with a compound key. 

The dataset is comprised of examples that represent attributes from a year and zipcode (I've attached a screenshot of part of the dataset). I would eventually like to run a time series analysis on it. 

 

So far I've tried using the set role operator to define two ID roles, but only one gets set...I've also tried importing the dataset again, but when I try to define more than one ID role, I am met with an error (that this role has already been taken). Please let me know if you have any suggestions for me. Thank you much.

 

 

Telcontar120

Yes, RapidMiner won't allow you to set more than one id attribute at the same time.  One simple fix is to create a new attribute which concatenates your two separate id variables and then assign that as the id using "Set Role".