Training a text classification model

Learn to train text classification models in IBM RPA Studio.

Before you begin

  • You need access to IBM RPA Studio.
  • You need training data. Make sure to read Planning for classifying text section to understand how to setup your training data and learn other concepts about text classification in IBM RPA.
Important:If you try to create a new model version for an existing text classification model, the new model always overwrites the previous model including the algorithm if you change it. This means that to improve an existing model, you always need to use your full training data instead of using only additional data.

Procedure

  1. Log in to IBM RPA Studio and select the tenant you want to use.
  2. In the menu bar on the header, select the Tools tab.
  3. Click the Machine Learning Model Builder button.
  4. In the Name field, type in a unique name.
    • Type a unique name to create a new text classification model.
    • Type an existing text classification model name to overwrite that model.
  5. In the Culture field, select the language the model supports.
  6. In the Algorithm field, select the algorithm that you want to use. See Text classification algorithms for details.
  7. In the Folder field, select the root folder containing the training data set.
  8. Optional: Select the Set as production checkbox to set the model as the production version.

Results

Your text classification model is available in the tenant you are logged in IBM RPA Studio. You can test your model in IBM RPA Control Center by following the next procedure:

  1. Log in to IBM RPA Control Center.
  2. Click Machine learning.
  3. Locate your model, click the ellipsis button ⋮, and select Test classifier.

What to do next

See Using a text classification model to learn how to use the trained text classification model.