Apply a R-Based model
This Action works with any partition type. This means that, using this Action, you can score any dataset size, including the datasets that are larger than your central RAM memory. This allow to by-pass the limitations on the RAM memory that all other tools that are using R as back-end exhibits. In this latest release, the objective of the model is saved in the file, and it is possible to open many models at the same time. In a single operation, you can score dozens of models of different types.
•Partition: set partition or partition variable if you have a few 100 thousands or millions of rows.
•Key Variable: Unique identifyer of each row
•R Model(s) to apply: select the models manually. This is not necessary if you use the input pin, which is recommended, as in the example below, in which we simply list *.rModel, filter those that we want to use, and apply on the new dataset.
•Include probability per class: Some algorithms allow returning probability additional to class. Select this option to return it.