"Time Series Prediction"

Hi!

Currently I'm using a process in Rapid Miner to predict the closing value of a specific stock of the subsequent day.

I feed the process with untransformed data, that means: (1) Date (as ID) (2) Closing value (as Attribute (Real)).


Do you think executing a data transformation, before does make sense in this case ? If so, which one should I choose and how exactly do I implement it?  ???


Best regards !