US 12,147,912 B2
Predicting a time of an event associated with an instrument using a machine learning model
Abdelkader M'hamed Benkreira, Brooklyn, NY (US); Joshua Edwards, Philadelphia, PA (US); and Michael Mossoba, Great Falls, VA (US)
Assigned to Capital One Services, LLC, McLean, VA (US)
Filed by Capital One Services, LLC, McLean, VA (US)
Filed on Aug. 17, 2021, as Appl. No. 17/445,253.
Prior Publication US 2023/0057762 A1, Feb. 23, 2023
Int. Cl. G06Q 30/02 (2023.01); G06N 5/04 (2023.01); G06N 20/00 (2019.01); G06Q 20/40 (2012.01); G06Q 30/0201 (2023.01); G06Q 40/02 (2023.01); G06Q 20/04 (2012.01)
CPC G06N 5/04 (2013.01) [G06N 20/00 (2019.01); G06Q 20/4037 (2013.01); G06Q 30/0201 (2013.01); G06Q 40/02 (2013.01); G06Q 20/0425 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A system for predicting a time of an event associated with an instrument using a machine learning model, the system comprising:
one or more memories; and
one or more processors, communicatively coupled to the one or more memories, configured to:
train a neural network model using historical data associated with historical transaction events relating to transactions involving instruments, the historical data for each historical transaction event indicating at least:
a time duration between provisioning of a respective instrument to a respective recipient and a respective transaction event associated with the respective instrument, and
a transaction type associated with the respective instrument, wherein the neural network model is trained to output one or more predictions of times of transactions events associated with instruments based at least on a transaction type associated with an instrument;
receive information indicating that the instrument has been provided by a user to a recipient,
wherein the instrument is associated with a value, and
wherein the instrument is designated for the recipient;
process an annotation of the instrument to determine a relationship associated with the user and the recipient,
determine a transaction type associated with the instrument based on processing the relationship and the annotation;
identify one or more characteristics associated with the recipient,
determine, using the trained neural network model, a prediction of the time of the event associated with the instrument based on the one or more characteristics associated with the recipient and the determined transaction type associated with the instrument;
determine, based on the prediction of the time of the event and the value, whether a balance of an account of the user will reach a threshold as a result of the event associated with the instrument; and
transmit, to a device of the user, a notification based on a determination that the balance will reach the threshold as the result of the event associated with the instrument.