Group Models (RapidMiner Studio Core)

Synopsis

This operator groups the given models into a single combined model. When this combined model is applied, it is equivalent to applying the original models in their respective order.

Description

The Group Models operator groups all input models together into a single combined model. This combined model can be applied on ExampleSets (using the Apply Model operator) like other models. When this combined model is applied, it is equivalent to applying the original models in their respective order. This combined model can also be written into a file using the Write Model operator. This operator is useful in cases where preprocessing and prediction models should be applied together on new and unseen data. A grouped model can be ungrouped with the Ungroup Models operator. Please study the attached Example Process for more information about the Group Models operator.

Input

model in (Model)
This input port expects a model. This operator can have multiple inputs but it is mandatory to provide at least two models to this operator as input. When one input is connected, another model in port becomes available which is ready to accept another model(if any). The order of models remains the same i.e. the model supplied at the first model in port of this operator will be the first model to be applied when the resultant combined model is applied.

Output

model out (Grouped Model)
The given models are grouped into a single combined model and the resultant grouped model is returned from this port.

Tutorial Processes

Grouping models and applying the resultant grouped model

The 'Iris' data set is loaded using the Retrieve operator. A breakpoint is inserted here so that you can have a look at the ExampleSet. You can see that the ExampleSet has four regular attributes. The Split Data operator is applied on it to split the ExampleSet into training and testing data sets. The training data set (composed of 70% of examples) is passed to the SVD operator. The dimensionality reduction and dimensions parameters of the SVD operator are set to 'fixed number' and 2 respectively. Thus the given data set will be reduced to a data set with two dimensions (artificial attributes that represent the original attributes). The SVD model (model that reduces the dimensionality of the given ExampleSet) is provided as the first model to the Group Models operator. The Naive Bayes operator is applied on the resultant ExampleSet (i.e. the training data set with reduced dimensions). The classification model generated by the Naive Bayes operator is provided as the second model to the Group Models operator. Thus the Group Models operator combines two models SVD dimensionality reduction model Naive Bayes classification model. This combined model is applied on the testing data set (composed of 30% of the 'Iris' data set) using the Apply Model operator. When the combined model is applied, the SVD model is applied first on the testing data set. Then the Naive Bayes classification model is applied on the resultant ExampleSet (i.e. the testing data set with reduced dimensions). The combined model and the labeled ExampleSet can be seen in the Results Workspace after the execution of the process.