US 12,423,623 B2
Systems and methods for using a predictive engine to predict failures in machine-learning trained systems for display via graphical user interface
Jason Hoover, Grapevine, TX (US); Micah Price, Plano, TX (US); Stephen Wylie, Carrollton, TX (US); Qiaochu Tang, Frisco, TX (US); and Geoffrey Dagley, McKinney, TX (US)
Assigned to Capital One Services, LLC, McLean, VA (US)
Filed by Capital One Services, LLC, McLean, VA (US)
Filed on Apr. 30, 2024, as Appl. No. 18/650,609.
Application 18/650,609 is a continuation of application No. 16/659,960, filed on Oct. 22, 2019, granted, now 12,001,927.
Prior Publication US 2024/0281715 A1, Aug. 22, 2024
This patent is subject to a terminal disclaimer.
Int. Cl. G06Q 50/18 (2012.01); G06N 5/04 (2023.01); G06N 20/00 (2019.01)
CPC G06N 20/00 (2019.01) [G06N 5/04 (2013.01); G06Q 50/188 (2013.01)] 9 Claims
OG exemplary drawing
 
1. A computer system for using a machine-learning predictive engine to predict failures in transaction occurrences, the computer system comprising:
at least one memory storing instructions; and
one or more processors configured to execute the instructions to perform operations including:
receiving electronic transaction data comprising an item and candidate transaction terms, the candidate transaction terms comprising a first transaction term, the electronic transaction data being associated with a candidate transaction involving a plurality of parties;
determining a first likelihood that all of the parties in the plurality of parties will agree to the candidate transaction terms based on the electronic transaction data by evaluating the candidate transaction terms with the machine-learning predictive engine, wherein the machine-learning predictive engine is configured to:
represent electronic prior transaction data of each of a plurality of prior transactions that occurred and each of a plurality of prior candidate transactions that did not occur as a respective vector in an n-dimensional space;
represent the electronic transaction data as a further feature vector;
perform a clustering operation to determine distances between the further feature vector and each of a first cluster corresponding to the plurality of prior transactions that occurred and a second cluster corresponding to the plurality of prior transaction that did not occur; and
determine the first likelihood based on the determined distances;
displaying an indication of the first likelihood that all of the parties will agree to the candidate transaction terms on a graphical user interface;
determining whether the first likelihood is below a predetermined threshold;
in response to determining that the first likelihood is below the predetermined threshold, determining one or more alternate terms of the candidate transaction that increase the first likelihood to a second likelihood that all of the parties will agree to the candidate transaction terms that is above the predetermined threshold, the determining based on identification of a modification to the further feature vector that would move the further feature vector closer to the first cluster; and
causing the graphical user interface to display the one or more alternative terms.