US 12,287,815 B2
System and method for generating a research reporting narrative
Ian Ratner, Silver Spring, MD (US); and Ettie Etela, Silver Spring, MD (US)
Assigned to Discovery Communications, LLC, Silver Spring, MD (US)
Filed by DISCOVERY COMMUNICATIONS, LLC, Silver Spring, MD (US)
Filed on Feb. 24, 2021, as Appl. No. 17/183,906.
Prior Publication US 2022/0269705 A1, Aug. 25, 2022
Int. Cl. G06F 16/332 (2019.01); G06F 16/31 (2019.01); G06F 16/3329 (2025.01); G06F 16/38 (2019.01); G06N 20/00 (2019.01)
CPC G06F 16/3329 (2019.01) [G06F 16/316 (2019.01); G06F 16/3322 (2019.01); G06F 16/38 (2019.01); G06N 20/00 (2019.01)] 17 Claims
OG exemplary drawing
 
1. A computer-implemented method, comprising:
receiving, by one or more processors, performance metric data from one or more server systems, wherein the one or more server systems are configured to utilize one or more data providers to collect the performance metric data corresponding to user associated content;
generating, by the one or more processors, via one or more trained machine-learning models, natural language data based on the performance metric data;
receiving, by the one or more processors, user query data from a user device corresponding to a user, wherein the user query data includes a user domain and a time range;
analyzing, by the one or more processors, the user query data to determine at least one key metric;
utilizing, by the one or more processors, via the one or more trained machine-learning models, the performance metric data to determine at least one driver associated with the at least one key metric;
selecting, by the one or more processors, via the one or more trained machine-learning models, a natural language data subset based on the at least one key metric, wherein the natural language data subset includes non-standardized data based on one or more collection processes of the one or more data providers corresponding to the at least one driver;
receiving, by the one or more processors, additional metric data from one or more domains based on the user domain and the time range;
aggregating, by the one or more processors, the additional metric data by applying domain knowledge from the one or more domains to the additional metric data;
analyzing, by the one or more processors, via the one or more trained machine-learning models, the performance metric data and the aggregated additional metric data to determine at least one metric update;
creating, by the one or more processors, via the one or more trained machine-learning models, contextual data corresponding to the at least one metric update;
converting, by the one or more processors, the natural language data subset to a standardized natural language data subset for updating based on the contextual data, wherein the natural language data subset is converted to a standardized format based on audience data associated with the user query data;
updating, by the one or more processors, the standardized natural language data subset based on the contextual data corresponding to the at least one metric update; and
transmitting, by the one or more processors, the updated standardized natural language data subset to the user device.