TVT (Train-Validate-Test) Split

Splits a single input dataset into training, validation, and testing subsets. The proportion of the training set is controlled by the user; the remaining data is split equally into the validation and testing sets.

The node also allows for stratified sampling based on a string or boolean column of the input data. The column contents must have a domain defined on them.

A seed value can be supplied for randomized sampling to allow for consistent results. If no seed is given, a random seed will be used each time.

Requires the "KNIME Streaming Execution" extension.

Input ports

1. Type: Table
   Input
   Input dataset

Output ports

1. Type: Table
   Training Set
   Training set. The size of this table is defined by the slider control in the configuration dialog.
2. Type: Table
   Validation Set
   Validation set. The size of this table is half of the remaining data after the training set has been split off.
3. Type: Table
   Testing Set
   Testing set. The size of this table is half of the remaining data after the training set has been split off.