Skip to main content


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

Monitoring text predictions

Updated on July 5, 2022

Explore how successful your text predictions are in predicting topics, entities, and sentiments in the messages that come through your conversation channels, such as, emails and chats. Gain insights by reviewing the statistics for your text predictions.

Performance measures can help you decide whether you need to create new topics or entities, or add more training data to train the models behind the text prediction.
  1. Open the text prediction:
    1. In the navigation pane of App Studio, click Channels.
    2. In the Current channel interfaces section, click the icon that represents a channel for which you want to configure the text prediction.
    3. On the channel configuration page, click the Behavior tab, and then click Open text prediction.
  2. On the Analysis tab, select the outcome that you want to analyze:
    • To view a summary for all the outcomes, click Overview.
    • To view the statistics for topic detection, click Topics.
    • To view the statistics for sentiment detection, click Sentiments.
    • To view the statistics for entity detection, click Entities.
    Result:

    Each tab contains charts that show performance measures that are relevant to the selected scope. For more information about the charts that are used for analysis, see Text prediction analysis charts.

  3. In the Language field, select the language settings:
    • To see the results for a specific language, select that language.
    • To see the results for all languages that are configured in the prediction settings, select All.
  4. In the Time frame field, select the period that you want to analyze.

    The following periods are available:

    Last 7 days
    7 days of daily data.
    Last 30 days
    30 days of daily data.
    Previous month
    Daily data for the previous month. For example, if today is 15 March 2021, then the charts show the daily data from 1 February 2021 to 28 February 2021.
    Last 3 months
    Last 13 weeks of weekly data. The charts show 13 dates and each date corresponds to the Sunday of that week. All days that follow a Sunday are captured in the data for that Sunday. For example, the data for 1 March and 6 March 2021 is aggregated and stored in the 28 February date as weekly data.
    Last 12 months
    Last 12 months of monthly data. The charts show 12 dates and each date corresponds to the first day of the month. For example, the data for 5 March and 15 March is aggregated and stored in the 1 March date as monthly data.
    Selecting a time frame for a text prediction analysis
    Time frame selection on the Analysis tab in a text prediction
What to do next: To improve performance, change the configuration of your prediction, for example, by creating new topics and entities, or adding more training data to train the models. For more information, see Analyzing messages with text predictions.
  • Text prediction analysis charts

    Learn how to understand the performance charts on the Analysis tab of a text prediction. Monitoring predictions is a source of valuable information about the messages that come through your conversational channels.

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