"Cross validation for regression"

lmsasulmsasu Member Posts: 20 Contributor II
edited May 2019 in Help
Hello,

how could I assess the performance of a regression system by cross validation? the x-validation operator seems to work only for classification. When I try to use it with a neural network inside, I get the error message: "X-Validation cannot handle numerical label".

Thanks

Answers

  • awchisholmawchisholm RapidMiner Certified Expert, Member Posts: 458 Unicorn
    Hello

    Change the sampling type to "shuffled sampling" in the X-Validation operator. The default "stratified sampling" keeps the proportion of classes the same for each sample it creates. If there are too many classes because the label is a continuous value, the operator will struggle because it will never be able to find exact matches for all examples in each sample.

    regards,

    Andrew
Sign In or Register to comment.