US 12,277,137 B2
Framework for analyzing, filtering, and prioritizing computing platform alerts using feedback
Akshay Arora, Delhi (IN); Krishna Mohan Roy, Varanasi / Uttar Pradesh (IN); Joydeep Dam, Bengaluru (IN); and Jayant Pimpalkar, Prades (IN)
Assigned to Accenture Global Solutions Limited, Dublin (IE)
Filed by Accenture Global Solutions Limited, Dublin (IE)
Filed on Apr. 12, 2023, as Appl. No. 18/133,569.
Prior Publication US 2024/0346037 A1, Oct. 17, 2024
Int. Cl. G06F 16/26 (2019.01); G08B 29/18 (2006.01); G06F 16/00 (2019.01)
CPC G06F 16/26 (2019.01) [G08B 29/185 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method for tracking and prioritizing observation data across multiple observational hierarchies in a computing platform, the method comprising:
obtaining one or more alerts and feedback indicators corresponding to at least one alert of the one or more alerts, each of the one or more alerts representing a portion of the observation data in a computing system, the computing system connected to the computing platform comprising a data mining engine, and each of the feedback indicators indicating a positive or negative association of a respective alert of the at least one alert for the represented portion of the observation data, wherein the observation data includes at least one of data requests, latency, traffic, processing loads, server requests;
identifying, based on the feedback indicators corresponding to the at least one alert, a first subset of the one or more alerts that correspond to a subset of the observation data with negative feedback;
determining a first set of alert attributes that include metadata related to the subset of the observation data for the first subset of the one or more alerts;
determining, based on the feedback indicators and by a data mining engine, a model from a plurality of models of the data mining engine that is configured to analyze a respective alert attribute of the first set of alert attributes for the observation data represented by the one or more alerts;
analyzing, by the model of the data mining engine, the first set of alert attributes to identify a subset of the first set of alert attributes with corresponding likelihoods representing that one or more alert attributes of the first set of alert attributes caused the negative association of the respective alert represented by the respective portion of the observation data;
filtering, based on the subset of alert attributes and by the data mining engine, the one or more alerts to exclude a second subset of the one or more alerts with the highest likelihood of the negative association; and
providing, for output of the computing platform, the one or more alerts that exclude the second subset of the one or more alerts.