US 12,124,835 B2
Computer system and method for facilitating real-time determination of a process completion likelihood
Francisco Javier Vicente Gonzalez, London (GB); Abhishek Singh, Maharashtra (IN); and Jonathan Owen, Manchester (GB)
Assigned to THE BANK OF NEW YORK MELLON, New York, NY (US)
Filed by THE BANK OF NEW YORK MELLON, New York, NY (US)
Filed on Nov. 2, 2022, as Appl. No. 17/979,353.
Claims priority of application No. 202241025073 (IN), filed on Apr. 28, 2022.
Prior Publication US 2023/0350662 A1, Nov. 2, 2023
Int. Cl. G06F 8/65 (2018.01); G06F 11/34 (2006.01)
CPC G06F 8/65 (2013.01) [G06F 11/3495 (2013.01)] 21 Claims
OG exemplary drawing
 
1. A method for facilitating real-time determination of a process completion likelihood, implemented on one or more processors configured to execute computer program instructions stored in memory, the method comprising:
obtaining first data comprising a first update to a first system, the first update occurring at a first time, wherein updates to the first system are permitted until an expiration time;
at least one of retrieving or receiving a set of fixed descriptors of the first system;
obtaining a set of one or more status updates describing the first system at one or more times prior to the first time, wherein each status update of the set of one or more status updates includes at least (i) an update to the first system and (ii) a time that the respective status update occurred;
selecting a trained machine learning model from a model database storing trained machine learning models, wherein the trained machine learning model comprises an extreme random tree model;
retrieving historical failure/success data of a set of systems from a training data database, wherein the historical failure/success data comprises, for each system of the set of systems, a respective set of fixed descriptors and a respective set of one or more status updates,
wherein the respective set of one or more status updates includes an indication of an update to a respective system of the set of systems and a timestamp associated with the update;
training the extreme random tree model using at least a portion of the historical failure/success data to obtain a trained extreme random tree model;
computing, based on the first data, the set of fixed descriptors, and the set of one or more status updates, using the trained machine learning model, a first failure/success score indicating a likelihood that, at the expiration time, the first system satisfies a threshold condition;
storing the first failure/success score in memory in association with the first data, the set of fixed descriptors, and the set of one or more status updates; and
generating a notification based at least in part on the first failure/success score.