Medical dataset

[ekyy2]

Hi,

I believe there is a mix up in the names of the medical train and test datasets. The train dataset is supposed to consist of 645 examples, while the test dataset is made up of 333 examples. Note however that there is a slight difference in the relation where the original test file has the invertSelection flag (-V) while the train file does not:

@relation 'MEDC: -C 45 -I 0-weka.filters.unsupervised.attribute.Remove-R1-45_MEDC: -C 45 -I 0-weka.filters.unsupervised.attribute.Remove-R46-1494-weka.filters.unsupervised.instance.RemovePercentage-P66-V'

Thank you.

Regards,
Edward

medical.zip