Skip to main content


         This documentation site is for previous versions. Visit our new documentation site for current releases.      
 

Creating a taxonomy for keyword-based topic detection

Updated on July 5, 2022

After you create a topic model, define the topics that you want to detect in a piece of text. For each topic, add a list of keywords that define the topic. Based on these keywords, topic detection then assigns topics to an analyzed piece of text.

Before you begin: Create a keyword-based topic model by specifying the model name, language, and corresponding ruleset. For more information, see Setting up a keyword-based topic model.
  1. In the Taxonomy workspace, create a list of topics that you want to detect:
    1. Create a parent topic by clicking Add topic.
    2. Optional: To add a child topic, select a parent topic, and then click ManageAdd Child.
      You can add multiple levels of topics, depending on your use case and classification approach. For example, you can break down the parent category Support into In-store support and Phone support.
    3. Repeat steps 1.a through 1.b to create a complete hierarchy of the topics that you want to detect.
  2. For each topic, enter a list of keywords that pertain to that topic.
    Tip: A keyword can consist of multiple words. To separate keywords, press the Tab+Enter keyboard shortcut.
    You can specify the following types of keywords:
    Should words
    If any of the Should words appear in a piece of text, topic detection assigns that text to the corresponding topic. To achieve accurate results, create an exhaustive list of Should words. For example, for a Support topic, you can specify the following Should words: help, assistance, support, aid, guidance, assist, advice, and so on.
    Must words
    If all Must words appear in a piece of text, topic detection assigns that text to the corresponding topic. You can specify whether you want all Must words to appear at sentence level, or in the text as a whole. Use Must words to narrow down your topic detection conditions. For example, you can specify that a piece of text must contain the word help to be assigned to the Support parent category.
    And words
    If a piece of text contains both And words and Should words, topic detection assigns that text to the corresponding topic. Use And words to distinguish between similar categories and to increase the accuracy of topic detection. For example, you can specify the same Should words for the In-store support and Phone support topics, but then add premises, store, and office as keywords specific to the In-store support topic, and phone and call as keywords specific to Phone support.
    Not words
    If a Not word appears in a piece of text, the text is not assigned to the corresponding topic. For example, enter phone or call as words that prevent topic detection from assigning a piece of text to the In-store support topic.
  3. To detect child topics only when the corresponding parent topic is detected, for the parent topic, select the Match child topics only if the current topic matches check box.
  4. Optional: To test your taxonomy, select ActionsTest.
    Tip: Always test your taxonomy on a number of text samples to determine whether it accurately assigns topics. Depending on the results, you might refine your taxonomy, for example, by increasing the number of Should words to accommodate additional use cases, or by adding Not words to help differentiate between similar categories.
  5. Optional: To export the taxonomy as an .xlsx file, select ActionsExport.
  6. Save the taxonomy by clicking Save.
    You can use the taxonomy as part of a machine learning topic model or directly in Text Analyzers to perform keyword-based topic detection.
Result: Your model is now available in the Models workspace.
  • Previous topic Setting up a keyword-based topic model
  • Next topic Importing a taxonomy for keyword-based topic detection

Have a question? Get answers now.

Visit the Support Center to ask questions, engage in discussions, share ideas, and help others.

Did you find this content helpful?

Want to help us improve this content?

We'd prefer it if you saw us at our best.

Pega.com is not optimized for Internet Explorer. For the optimal experience, please use:

Close Deprecation Notice
Contact us