US 12,273,587 B2
Use case configuration system and method
Richard Dowling, Glasgow (GB); Bryan Bossard, Canyon Country, CA (US); Peter Docherty, Glasgow (GB); and Christopher McGuire, Glasgow (GB)
Assigned to ThinkAnalytics Ltd., Glasgow (GB)
Filed by ThinkAnalytics Ltd., Lanarkshire (GB)
Filed on Apr. 27, 2023, as Appl. No. 18/140,099.
Prior Publication US 2024/0364965 A1, Oct. 31, 2024
Int. Cl. H04N 21/442 (2011.01); H04N 21/258 (2011.01)
CPC H04N 21/44213 (2013.01) [H04N 21/25891 (2013.01); H04N 21/44204 (2013.01)] 17 Claims
OG exemplary drawing
 
1. A method for determining use cases for a content selection interface that provides user selectable indications of content available from a content provider according to the use cases, the indications of available content being selectable by a user to have the content provided to the user by the content provider, the method comprising, using a computing system:
accessing, from at least one computer readable data store, historical data representing previous user actions of a plurality of users using a content selection interface;
identifying, using the computing system, one or more attributes associated with the user actions from the historical data;
using the identified attributes to configure one or more use cases for providing indications of available content on the content selection interface;
identifying data from the historical data for a first time period and data from the historical data for a second time period, wherein the first time period precedes the second time period, and the configuring of the use cases comprises using the use cases identified to be associated with at least some of the user actions of the first time period to configure use cases or a content recommendation engine for providing indications of available content on the content selection interface, and
using at least some of the user actions of the second time period to assess performance of the configured use cases, wherein the data is date and/or time stamped and the method comprises performing a wind-back to only use data, parameters and content recommendation models dated and/or time stamped for a date and/or time during the first time period.