Split Dataset for Validation

KNIME Analytics Platform
1

Hello,

I have been working with a decent sized dataset (167k rows).  I have done all the transformations necessary to use LibSVM however I don't know how to split my dataset in KNIME to have ~80% of it used for training and ~20% of it used for validation.  Is there an easy way to do this?

Regards,

SC

2

Just use the Partitioning node or the mode sophisticated Cross Validation meta node.

3

Hi,

Try: 'Data manipulation' -> 'Row' -> 'Transform' -> 'Partitioning'.

If you would like to do some cross validation, then try: 'Mining' ->'Meta' -> 'Cross validation'

Regards

4