Creating keyword-based text extraction models

A keyword-based text extraction model contains a user-defined list of keywords that are usually related to a specific domain. For example, the list can contain names of companies that are similar to yours and that do business in a specific location.
Before you begin: Make sure that the system locale language settings are set to UTF-8.
  1. In the navigation panel of Prediction Studio, click Predictions.
  2. In the header of the Predictions work area, click New > Text extraction.
  3. In the Create Text Extraction Model window, enter the name of the text extraction model.
  4. In the Creation section, select Keyword list.
  5. In the Language section, expand the drop-down list and select a language for the model to use.
  6. Perform one of the following actions:
    • To manually create a list of keywords, select Create keyword list.
    • To upload a list of keywords from a file, click Upload a file.
  7. If you selected Create keyword list, perform the following actions:
    1. In the Word field, enter the keyword, for example, US president.
    2. In the Synonyms field, provide a comma-separated list of keyword synonyms.
      For example: New York – NY, NYC, Big Apple, The Five Boroughs
    3. Optional: Click Add item to define more keywords.
  8. If you selected Upload file, perform the following actions:
    1. Optional: Download the template that contains the required structure and examples by clicking Download template.
      You can edit that file to provide your own keywords.
    2. Click Choose file to select and upload a file with keywords from your directory.
      To delete obsolete keywords from the model, clicking the Trash can icon.
  9. Optional: Click Add item to add more keywords to the model.
  10. In the Save model section, finalize the creation of a keyword-based text extraction model by providing its application context:
    • To use the default rule context for decision data rules that contain sentiment analysis models, select Use default context.
    • To manually specify the Applies to class, ruleset, and ruleset version parameters of the new rule, select Specify context.
  11. Click Create.
What to do next:  Test your model on real-life data.