"How can X-Validation use only a training set and give a performance?"
I There I'm a rookie to Rapidminer but there are something that is bugging me.
When we use X-Validation its only required the training set, and in the test side we apply the model and test performance
but how could we test performance if there arent any test data?
Does X-Validation train itself with the training set (labeled) and than unlabel the same training set and test its performance?
Thanks a Lot guys
When we use X-Validation its only required the training set, and in the test side we apply the model and test performance
but how could we test performance if there arent any test data?
Does X-Validation train itself with the training set (labeled) and than unlabel the same training set and test its performance?
Thanks a Lot guys
Find more posts tagged with
Sort by:
1 - 2 of
21
Thanks a lot Andrew, Quite Explanative 
Cross validation splits the training data into N partitions and builds a model on N-1 of these to apply to the 1 partition that is left over to create a performance measure. It repeats this N times for the N different individual partitions to obtain N performance measures which it averages. The result is an estimate of how the model would perform on unseen data.
Andrew