PREDICTING RAINFALL USING HISTORY DATA (AUTOMODEL)

Matt98Matt98 Member Posts: 1 Newbie
Hi. Im still new to the rapidminer and currently, im using automodel in rapidminer to predict 30 fears future rainfall using 30 years history data. As im running the model, the result that i get only 40% of all result. Overall data that i have is 14245 and the prediction that i get only  5698. From the rapidminer description, it said that the rapidminer will show only 40% of hold out set. So, what can i do to see all the data which is 100% of the prediction before i export it to the excell file?.



Answers

  • BalazsBaranyBalazsBarany Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert Posts: 835 Unicorn
    Hi!

    It's not a good practice to predict on data that went into the model. So AutoModel doesn't.

    It does a "split validation", splitting up the data into a training set (for building the model) and a holdout or test set (for validation).

    AutoModel tells you the best model and gives you a prediction process. If you change the split validation in this process to a cross validation (or build a new, simpler process with the same settings as the best model), you can get predictions for all examples. 

    Regards,
    Balázs

    Matt98
Sign In or Register to comment.