Skip to main content

Smart Search Features

edward_a
  • Updated

Popular Searches

Displays frequently searched terms that successfully returned results, ranked by the number of searches in descending order.

Refer to the developer guides for detailed instructions on setting up and configuring popular searches.

Popular Searches via Reference Files

Provides the ability to seed or hard-code search terms to pre-emptively guide user behaviour.

Steps to Configure:

  • Access Data section and then Datasets.

  • Create new Dataset and choose Popular Searches as a Template, give it a Name and Description

  • Save Dataset then open it again and edit through Editor or Upload a file with supported schema.

Key Attributes in the Dataset:

  • Search: Represents the search term as a user would input it.
  • Support: Determines the ranking or order of the search term within the suggestions. Higher support values ensure the term appears more prominently.
  • Constant: A boolean flag that controls the behavior of the search term:
    • TRUE: The term always appears at the top of the suggestion list, regardless of actual query matches.
    • FALSE: The term is treated as a normal suggestion and may be overridden by real-time search results.

Additional Documents in Search

Allows linking external resources like manuals or FAQs to enhance the context of search results.

Steps to configure:

  • Access Data section and then Datasets.

  • Create new Dataset and choose Additional documents as a Template, give it a Name and Description

  • Save Dataset then open it again and edit through Editor or Upload a file with supported schema.

Key Attributes in the Dataset:

Define each document by filling in the relevant attributes:

  • Type - category for grouping purposes. In a Search widget displayed as a tab.
  • Document - represents the search term as a user would input it.
  • Text - title for sub category.
  • Description - internal description.
  • Blurb - clickable title of a specific document.
  • URL - to direct users to the document when clicking on blurb.

How the results will look like:

Gaming Synonyms

The Synonym Mapping dataset allows you to link search terms that are treated equivalently in search results. For example, a search for "bacbao" will return the same results as "bac bo" and vice versa. Also could be used to define translations.

Steps to configure:

  • Access Data section and then Datasets.
  • Create new Dataset and choose Additional documents as a Template, give it a Name and Description

  • Save Dataset then open it again and edit through Editor or Upload a file with supported schema.

Key Attributes in the Dataset:

  • Synonyms - list synonyms, separating by | sign, e.g. bacbao|bac bo

How the results will look like:

Gaming Search Recommendations

The Search Recommendations feature provides personalized search results based on machine learning models.

This functionality is configured by the Opti-X team and must be requested for activation.

  • Models Used: The recommendations are powered by two types of models:

    1. Collaborative Filtering (Similar Users): Suggests items based on the preferences of similar users.
    2. Content-Based Filtering (Behavioral): Recommends items based on user behavior and preferences.

    You can also opt for a hybrid model combining both approaches by request.

    The system does not rely on bet_history for recommendations. Instead, search results from the user's input query are used to generate a "profile" that helps to score and recommend other relevant items.

  • How It Works:

Key Considerations: There should be no overlap between the items returned in the search results and the items generated by search recommendations. The two data sets remain separate to ensure the relevance of recommendations without duplicating search results.

Gaming Tag-Based Filtering

Refer to the Integration Guides to configure this feature.

Region-Based Filtering

Refer to the Integration Guides to configure this feature.

Sports Synonyms

The Synonym Mapping dataset allows you to link search terms that are treated equivalently in search results. For example, a search for "bacbao" will return the same results as "bac bo" and vice versa. Also could be used to define translations.

Steps to configure:

  • Access Data section and then Datasets.
  • Create new Dataset and choose Synonyms as a Template, give it a Name and Description

  • Save Dataset then open it again and edit through Editor or Upload a file with supported schema.

Key Attributes in the Dataset:

  • Synonyms - list synonyms, separating by | sign, e.g. man|manchester
  • Level:
    • 1 - Matched Events (Level 1): Synonyms are mapped to events (e.g., "Champions League" = "UCL"), returning results related to the event.
    • 0 - Matched Events & Matched Bets (Level 0): Synonyms apply to both events and associated betting options, such as betting odds or markets tied to the event (e.g., "Football World Cup" = "FIFA World Cup").

How the results will look like:

Level 1

Level 0

Sports Event Aliases

Event Aliases allows synonyms for events (e.g., "Real Madrid vs Barca" can be called "El Clasico").

Steps to configure:

  • Access Data section and then Datasets.
  • Create new Dataset and choose Event aliases as a Template, give it a Name and Description
  • Save Dataset then open it again and edit through Editor or Upload a file with supported schema.

Key Attributes in the Dataset:

  • From: Defines the event term used in the dataset (e.g., "Real Madrid vs Barca").

  • To: Defines how the synonym should appear in search requests (e.g., "El Clasico").

  • Infer Map: A legacy attribute, which is currently should be set to always true but will be removed soon.

     

    Synonyms do not need to match actual event names, just that both team names must appear in the event name.

Sports Entity Aliases

Entity Aliases works similarly to synonyms but gives more control over how synonyms are mapped to documents. It allows users to define how specific terms are treated for different entities.

Use Case Example: For instance, there are two football teams named "Liverpool" — one in the English Premier League and another in the Uruguayan Premier Division. The alias "The Reds" would apply only to the Liverpool team from the English Premier League, not the one from Uruguay.

Steps to configure:

  • Access Data section and then Datasets.
  • Create new Dataset and choose Event aliases as a Template, give it a Name and Description
  • Save Dataset then open it again and edit through Editor or Upload a file with supported schema.

Key Attributes in the Dataset:

  • Key Attributes:
    • From: The term that will be recognized (e.g., "Liverpool" or "The Reds").
    • To: Defines the alias or synonym that should be used in search requests (e.g., "Liverpool" = "The Reds").
    • Class Name and Type Name: These are optional fields used for grouping or categorizing entity aliases.
    • Infer Map: A legacy attribute, which is currently should be set to always true but will be removed soon.

How the results will look like:

Sports Price-Based Filtering

Refer to the Integration Guides to configure this feature.

Sports Time-Based Filtering

Refer to the Integration Guides to configure this feature.

Related articles