Applies a well-specified data transformation to a dataset
Category: Machine Learning / Score
Applies to: Machine Learning Studio
This content pertains only to Studio. Similar drag and drop modules have been added to the visual interface in Machine Learning service. Learn more in this article comparing the two versions.
This article describes how to use the Apply Transformation module in Azure Machine Learning Studio, to modify an input dataset based on a previously computed transformation.
For example, if you used z-scores to normalize your training data by using the Normalize Data module, you would want to use the z-score value that was computed for training during the scoring phase as well. In Azure Machine Learning Studio, you can do this easily by saving the normalization method as a transform, and then using Apply Transformation to apply the z-score to the input data before scoring.
Azure Machine Learning Studio provides support for creating and then applying many different kinds of custom transformations. For example, you might want to save and then re-use transformations that do the following:
Remove or replace missing values, using Clean Missing Data
Create a set of compact features by calculating joint probability distribution for a dataset, using the Learning with Counts modules.
How to use Apply Transformation
Add the Apply Transformation module to your experiment. You can find thi module under Machine Learning, in the Score category.
Locate an existing transformation to use as an input.
If the transformation was created earlier in the experiment (for example, as part of a cleaning or data scaling operation) typically the ITransform interface object is available on the module's right-hand output. Connect that output to the left-hand input of Apply Transformation.
Previously saved transformations can be found in the Transforms group in the left navigation pane.
If you design a transformation for an experiment but do not explicitly save it, the transformation is available in the workspace as long as your session is open. If you close the session but do not save the transformation, you can re-run the experiment to generate the ITransform interface object.
Connect the dataset that you want to transform. The dataset should have exactly the same schema (number of columns, column names, data types) as the dataset for which the transformation was first designed.
No other parameters need to be set; all customization is done when defining the transformation.
To apply a transformation to the new dataset, run the experiment.
To see how this module is used in machine learning, see the Azure AI Gallery:
Learning with Counts: Uses Apply Transformation to reuse a count table.
The Apply Transformation module can take as input the output of any module that creates an ITransform interface. These modules include:
|Transformation||ITransform interface||A unary data transformation|
|Dataset||Data Table||Dataset to be transformed|
|Transformed dataset||Data Table||Transformed dataset|
|Error 0003||Exception occurs if one or more of inputs are null or empty.|
For a list of errors specific to Studio modules, see Machine Learning Error codes.
For a list of API exceptions, see Machine Learning REST API Error Codes.