Categories

Versions

You are viewing the RapidMiner Studio documentation for version 10.1 - Check here for latest version

Get Parameters (Operator Toolbox)

Synopsis

This Operator extracts all parameters, including expert parameters, of another operator.

Description

This operator extracts all parameters, including expert parameters, of another operator (in the following called TargetOperator) . The name of the TargetOperator has to be specified. The parameters are converted into a Parameter Set. In addition a through port pair extender allows to through put any number of objects through the Get Parameters operator. Thus allows to control the process execution. The TargetOperator is received over the process context by its name. It is even possible to get the parameters of an TargetOperator not yet executed in the process. Be aware that the parameters are extracted at the execution time of the Get Parameters operator. In case of changing parameters be sure that Get Parameters is after the execution of the TargetOperator. The resulting Parameter Set can be stored in the repository, written to file with the Write Parameters operator, used by the Set Parameters operator or converted to an ExampleSet with the Parameter Set to ExampleSet operator of the Converters Extension (available at the Marketplace).

Currently the following features are not supported:

1. Macros are not evaluated. Instead the string %{<macro_name>} is taken as the value of the parameter. But the value of the Macro can easily be added for example as an additional attribute by the Generate Attribute operator.

2. The extraction of parameters within an parallel executed operator is not yet supported. Please disable parallel execution.

3. There are types of Parameters which are not converted into an easy readable String. Nevertheless they are written to the ParameterSet.

Input

  • through (IOObject)

    Any object you want to pass through the operator.

Output

  • parameters (Parameter Set)

    The Parameter Set containing the parameters extracted from the TargetOperator.

  • through (IOObject)

    The input object is passed through the operator and is delivered at the output port.

Parameters

  • operator_name The name of the TargetOperator which parameters are extracted. Range:

Tutorial Processes

Get parameters of decision tree

This tutorial process trains a Decision Tree and extracts all parameters of the Decision Tree operator with the Get Parameters operator.

Get all parameters of optimized model

This tutorial process retrieves the Golf data set. With an Optimized Parameters (Grid) operator a decision tree is trained. Inside the Optimize Parameters (Grid) operator the Get Parameters operator extracts the parameters of the final model and deliver the Parameter Set to the result port. The final parameters of the optimization are also delivered to the result port. Be aware that the Parameter Set of the Get Parameters operator contains all parameters of the trained decision tree whether or not they are part of the optimization process. The Parameter Set of the Optimize Parameters (Grid) operator contains only the parameters to be optimized. But here also parameters of other Operators within the Optimization could be included.

Train different learners and extract automatically their parameters

This tutorial process combines two operators of this extensions, the 'Extract Last Modifying Operator' operator and the 'Get Parameters' operator. Be aware that this tutorial process is an enhancement of the 'Train different learners and extract learner name' tutorial process of the 'Extract Last Modifying Operator' operator.

In this tutorial process the 'weighting' data set is splitted into a training and a test set. Within a loop, three different models are trained (selected by the 'Select Subprocess' operator). The name of the current model is automatically extracted by the 'Extract Last Modifying Operator' operator and used as an index attribute in the 'Generate Attribute' operator. In addition the name is used by the Get Parameters operator to extract all parameters of the current model. After the loop, a data set is created by the 'Pivot' operator which contains the confidences and predictions of the different models as additional attributes. In addition the different Parameter Sets for the different models are delivered in a collection to the result port. Be aware that the parallel execution of the Loop is disabled.