Table of contents

Extract features from text data in an SPSS Modeler flow

You can use the Features Extractor function in SPSS® Modeler flow.

This feature requires both the Watson Studio Local - SPSS Modeler Add On and Watson Machine Learning - Watson™ Explorer Add On.

  1. Open a project in the Watson Studio Local client.
  2. Add a Watson Explorer Collection that has annotators. For example, Part of Speech.
  3. Create an SPSS Modeler flow.
  4. Add the data from the dataset in the project and nodes to build a model.
  5. For text analysis, add the WEX Feature Extractor node.
    feature extractor
  6. Double-click the WEX Feature Extractor node to specify the properties. Under SETTINGS, specify the text fields for the feature extraction. The maximum number of features can be specified.
    node settings
  7. Select the collection name to use the feature extraction. When you select the collection name, you can see the categories for the features in the Feature Tree View. You need to select one or more categories.
  8. Build a model with the extracted features.
    spss flow

Execution of an SPSS Modeler stream with the WEX Feature Extractor node as a job

You can execute a flow with WEX Feature Extractor node as a job in the same manner as other flows. You can also run the job on Watson Machine Learning.