| CPC H04L 63/0853 (2013.01) [G16Y 30/10 (2020.01); H04L 63/107 (2013.01)] | 17 Claims |
|
1. A method for authenticating a financial transaction associated with a user using a plurality of IoT enabled devices, the method comprising:
receiving, by a transaction computing platform comprising at least one processor and memory, from the plurality of IoT enabled devices information regarding interactions with the user of the plurality of IoT enabled devices, wherein the information regarding interactions comprises geolocation information indicating physical locations of the interactions and time stamp information indicating times of the interactions;
receiving, by the transaction computing platform, additional interaction information from at least one third party IoT enabled device regarding interactions with the user, wherein the at least one third party IoT enabled device is operated by an entity distinct from the user and the transaction computing platform;
generating, by the transaction computing platform, a first constellation based on the received information regarding the interactions with the user from the plurality of IoT enabled devices, wherein the first constellation represents a pattern of the user's interactions with the plurality of IoT enabled devices over a first time period;
generating, by the transaction computing platform, a third constellation based on the received additional interaction information from the at least one third party IoT enabled device and the received information regarding the interactions with the user from the plurality of IoT enabled devices, wherein the third constellation represents a combined pattern of the user's interactions with both the plurality of IoT enabled devices and the at least one third party IoT enabled device over a second time period;
comparing, by the transaction computing platform, the first constellation to at least a second constellation stored in the memory of the transaction computing platform, wherein the second constellation represents a historical pattern of the user's interactions with the plurality of IoT enabled devices over a prior time period;
determining, by the transaction computing platform using a neural network, a constellation pattern from the compared first constellation and the at least the second constellation, wherein the constellation pattern represents a unique digital fingerprint of the user's habitual IoT interactions derived from the geolocation information and the time stamp information;
wherein determining, by the transaction computing platform using the neural network. the constellation pattern further comprises:
analyzing, by the neural network, temporal sequences of the interactions with the user from the plurality of IoT enabled devices based on the time stamp information;
correlating, by the neural network, the temporal sequences with the geolocation information to identify habitual spatial-temporal patterns of the user;
outputting, by the neural network, the constellation pattern based on the habitual spatial- temporal patterns
wherein the neural network comprises a Long Short-Term Memory (LSTM) neural network configured to process sequential data from the plurality of IoT enabled devices
determining, by the transaction computing platform, a trust status based on the determined constellation pattern, wherein the trust status indicates whether the financial transaction is authorized by the user;
determining, by the transaction computing platform, a risk ranking based on the determined constellation pattern, wherein the risk ranking quantifies a likelihood of fraud in the financial transaction based on deviations of the determined constellation pattern from the historical pattern represented by the second constellation;
updating, by the transaction computing platform, the risk ranking in real-time based on a defined time period and detection of a newly generated constellation, wherein the newly generated constellation is based on subsequently received information from the plurality of IoT enabled devices; and
transmitting, by the transaction computing platform, the trust status to the plurality of IoT enabled devices to enable approval or denial of the financial transaction based on the trust status.
|