Page tree
Skip to end of metadata
Go to start of metadata
The category boost feature is a dynamic feature that will generate a per-category feature automatically based on a Field Expression that produces an enum value, such as a category name or table name.

Category boosts are most useful for use when Training Relevancy Models, as this can automatically include categorical features for a document without needing to explicitly configure them and know them in advance.

See Configuring Relevancy Models for incorporating category boosts into manually configured relevancy models.

It is recommended that the field used for a category boost should not be overly specific or produce too many categories. If this is the case, it can result in category boosts dominating trained relevancy models, overshadowing phrase boosts that indicate the quality of the match for query terms.


A category boost has the following parameters:

nameThe name of the feature, for referencing in relevancy models.
fieldThe Field Expression that will specify the categories for a document, specified using REST syntax.
    <f:relevancyFeatures name="default">
      <f:categoryBoost name="table_" field="table"/>
    <xsd:complexType name="categoryBoost">
      <xsd:attribute name="name" type="xsd:string" use="required" />
      <xsd:attribute name="field" type="xsd:string" use="required" />

    Creating A New Category Boost

    To create a new category boost, click the Add a Feature button () at the top of the Relevancy Features page.

    This opens the Create Relevancy Feature wizard with two cards: Define and Type-Specific Properties.

      On the Define page, specify a name for the feature and select "Category Boost Feature" for the Feature Type.

      On the Type-Specific Properties page, specify the field that will indicate a document's category and click Create Relevancy Feature when done.

      See Managing Relevancy Features for more documentation on managing and working with relevancy features.

      • No labels