Back to results
US 12,705,568 B2Granted
Systems and methods for transaction acceptance modeling
AssigneeCapital One Services, LLC
InventorsWilson Kung, Mark Timms, Lenni Kimawati, Peter Deng
Life of this patent
Estimating the term…
Abstract
Systems and methods for determining the likelihood of a supplier accepting a certain transaction method such as a payment card. Exemplary systems can generate a machine learning predictive model to determine the likelihood of a supplier accepting a card payment.
Claims20
A system for generating an artificial intelligence model for predicting transaction acceptability, comprising:
- a memory for storing organized data comprising merchant and transaction data;
- at least one processor communicatively coupled to the memory and configured to:
- receive at least one payfile input;
- generate artificial intelligence model training data by creating a data package comprising supplier data enriched with card source data and supplier industry data, as well as supplier outreach results data from the at least one payfile input;
- set a plurality of model weights based on a plurality of connections between a plurality of hidden layers in the artificial intelligence model;
- analyze, via the artificial intelligence model, the artificial intelligence model training data with the plurality of model weights;
- adjust the plurality of model weights based on an output of the analysis;
- generate a finalized artificial intelligence model based on the output of the analysis and adjusted plurality of model weights;
- generate at least one transaction acceptability prediction with the finalized artificial intelligence model;
- and
- update the at least one transaction acceptability prediction without refitting the finalized artificial intelligence model by:
- generating a prior transaction acceptability score using the finalized artificial intelligence model,
- collecting updated supplier transaction acceptance information, and
- combining the prior transaction acceptability score with the updated supplier transaction acceptance information using Bayesian updating to generate an updated transaction acceptability score.
11 more elements
A method for generating an artificial intelligence model for predicting transaction acceptability, comprising:
- receiving, via a processor, at least one payfile input;
- generating, via the processor, artificial intelligence model training data by creating a data package comprising supplier data enriched with card source data and supplier industry data, as well as supplier outreach results data from the at least one payfile input;
- setting, via the processor, a plurality of model weights based on a plurality of connections between a plurality of hidden layers in the artificial intelligence model;
- analyzing, via the artificial intelligence model, the artificial intelligence model training data with the plurality of model weights;
- adjusting, via the processor, the plurality of model weights based on an output of the analysis;
- generating, via the processor, a finalized artificial intelligence model based on the output of the analysis and adjusted plurality of model weights;
- generating, via the processor, at least one transaction acceptability prediction with the finalized artificial intelligence model;
- and
- updating, via the processor, the at least one transaction acceptability prediction without refitting the finalized artificial intelligence model by:
- generating a prior transaction acceptability score using the finalized artificial intelligence model,
- collecting updated supplier transaction acceptance information, and
- combining the prior transaction acceptability score with the updated supplier transaction acceptance information using Bayesian updating to generate an updated transaction acceptability score.
9 more elements
A non-transitory computer-accessible medium having stored thereon computer-executable instructions wherein, when a computer hardware arrangement executes the instructions, the computer hardware arrangement is configured to perform procedures comprising:
- receiving, via a processor, at least one payfile input;
- generating, via the processor, artificial intelligence model training data by creating a data package comprising supplier data enriched with card source data and supplier industry data, as well as supplier outreach results data from the at least one payfile input;
- setting, via the processor, a plurality of model weights based on a plurality of connections between a plurality of hidden layers in the artificial intelligence model;
- analyzing, via an artificial intelligence model, the artificial intelligence model training data with the plurality of model weights;
- adjusting, via the processor, the plurality of model weights based on an output of the analysis;
- generating, via the processor, a finalized artificial intelligence model based on the output of the analysis and adjusted plurality of model weights;
- generating, via the processor, at least one transaction acceptability prediction with the finalized artificial intelligence model;
- and
- updating, via the processor, the at least one transaction acceptability prediction without refitting the finalized artificial intelligence model by:
- generating a prior transaction acceptability score using the finalized artificial intelligence model,
- collecting updated supplier transaction acceptance information, and
- combining the prior transaction acceptability score with the updated supplier transaction acceptance information using Bayesian updating to generate an updated transaction acceptability score.
9 more elements
Dependencies are inferred from the claim text.