CPC G06Q 40/02 (2013.01) [G06N 20/00 (2019.01); G06Q 40/12 (2013.12)] | 20 Claims |
1. A machine-learning method implemented by a computing system of a financial institution, the method comprising:
retrieving, by the computing system, one or more news alerts from one or more third-party devices;
applying, by the computing system, an entity extraction model to the one or more news alerts to identify one or more persons or organizations associated with the one or more news alerts;
applying, by the computing system, an entity resolution model to determine which of the one or more persons or organizations is a customer of the financial institution;
identifying, by the computing system, one or more topics of the one or more news alerts associated with the customer;
generating, by the computing system, an enriched alert comprising customer data and the one or more news alerts or a subset thereof, the enriched alert comprising a list of previous transactions involving the customer and a list of prior enriched alerts generated for the customer;
applying, by the computing system, a predictive model that receives the enriched alert and the one or more topics as input and generates (i) a probability that the enriched alert will trigger a suspicious activity report following an investigation, and (ii) a respective significance score for each of the one or more topics;
responsive to generating the probability, selecting, by the computing system, from a plurality of destination devices, based on the respective significance score for each of the one or more topics and based on the probability, a suitable destination computing device to which to transmit the enriched alert; and
transmitting, by the computing system, the enriched alert to the suitable destination computing device for investigation.
|