How to split any sizeable dataset into Training, Validation and Test sub-datasets using Spotfire

[Mohd Abid Dhamiri Bin Atan]

Just an arbitraryexample here... let's say I would like to splitthe Red Wine Quality dataset (https://www.kaggle.com/uciml/red-wine-quality-cortez-et-al-2009) into 3 datasets of 75%, 20% and 5% ratios for the purposes of Training, Validation and Testing of my model.

 

What are the different methods of achieving this in Spotfire I would like to purely rely on Spotfire without any R codes involved.

 

p.s I am not exactly using the Red Wine Quality dataset as mentioned above but my current dataset has similar number of variables and observations.

[Gaia Paolini]

Create an intermediate random column between zero and one:

[RANDOM01] : Rand(123)

Then since this will be (pseudo)-uniformly distributed, you can assign rows to a sample by:

[sAMPLE_ID]:

case

when [RANDOM01]