Venn-Abers prediction

A node for creating Venn-Abers predictors. It will automatically handle binary classification problems. Assign class attribute and the proportion of training instances to be used as calibration.

Version 0.1: Implemented classification

Input ports

1. Type: Table
   Training and calibration
   This data will be split into two fractions using the calibration input as fraction. The set not used for calibration is the proper trainign set, used to train the underlying model (a Random Forest). The calibration set is used to calibrate the conformal predictor.
2. Type: Table
   Test
   This set is solely used for testing. This set, together with added results, are also being fed out through the first output port.

Output ports

1. Type: Table
   Test with results
   The test set with added results.
2. Type: Table
   Port 2