PMML Ensemble Predictor

The PMML Ensemble Predictor node takes a pmml document containing an ensemble of other models and executes all of them on the given data. The results from all models are then aggregated using a method specified in the ensemble model. The output contains a column for each prediction of the single models (if set in the settings) and one column with the combined result. Note that ensembles of ensembles are not supported with the exception of Gradient Boosted Trees which are treated base model. It's on the other hand not possible to predict Gradient Boosted Trees models with this predictor, please use the PMML Predictor node or the Gradient Boosted Trees Predictor (PMML) nodes for this task.

Options

Return individual predictions

Determines whether the output contains one column for each model's prediction

Multiple model method as column name

Determines whether the column with the combined predictions should be named after the combination method

Tie break

Determines what is output when there is a tie between two classes in the combination methods majority vote and weighted majority vote

Views

This node has no views

Workflows

• 01_Model_Selection_SampledKNIME Hub

Links

• No links available

Developers