A machine-learning algorithm is trained with features relevant to basket data for items of transactions. The trained algorithm is trained to predict whether a given transaction is more or less likely to be associated with theft being engaged in by a transaction operator for the transaction. The trained algorithm is then provided basket data for a given transaction and produces as output a theft prediction value. When the theft prediction value exceeds a configured threshold value, the transaction is flagged for manual intervention or the transaction is flagged for subsequent manual verification.

Image analysis is performed on a user at a transaction terminal. Based on behaviors, expressions, and activities of the user, fraud or potential fraud is flagged. When fraud is flagged, the transaction terminal stops processing an active transaction on behalf of the user and alerts are sent. When potential fraud is flagged, images/video associated with the active transaction are sent for review and the active transaction may be suspended or permitted to proceed at the transaction terminal. In an embodiment, a same user conducting multiple transactions with different accounts at a same transaction terminal or multiple different transaction terminals within a configured period of time is automatically identified as fraud based on a fraud rule.

The invention relates to a method and system that combines payment data and cyber fraud indicators to identify potential fraud in payment requests from a client. The system comprises: a memory that stores and maintains a list of known fraud characteristics and cyber fraud indicators; and a computer processor, coupled to the memory, programmed to: receive, via an electronic input, a payment instruction from the client; identify one or more cyber fraud indicators associated with the payment instruction; apply payment decisioning to merge the one or more cyber fraud indicators to the payment instruction; generate a risk score based on the payment decisioning to determine whether the payment instruction should be executed; and automatically apply the payment decisioning to the payment instruction.

While mobile device location remains, private, user transaction controls are applied to a specific authorization request. The user transaction controls are pre-configured by the user of the mobile account holder device and identified by the enriched merchant data. Location algorithms predict transaction locations used to obtain enriched merchant data responsive to location privacy mode. Responsive to the user transaction controls, a fraud recommendation response is sent to the approval system, in real time with the transaction. The fraud recommendation response prevents a false negative by using an enriched merchant location rather than the raw merchant location.

A biometric matching process is disclosed. The biometric matching process may be used to obtain access to a resource managed by an access device using only biometric information. In some embodiments, a biometric template is stored in relation to a user device and/or account information, and is obscured. Upon receiving a request for access to a resource from an access device, the system may identify a number of user devices in proximity to the access device. Biometric templates associated with each of those user devices may be compared to a biometric template received from the access device. Upon identifying a match, the system may provide the access device with account information stored in relation to the matched biometric template. The access device may then complete a transaction using the provided account information and grant access to the requested resource.

Disclosed are systems, methods, apparatuses, and computer readable media for quickly and efficiently providing individual-level scores that are serve as more-accurate predictions of a target event. These predictions are made using a model that factors specific variables based on transaction attributes gathered from transaction data for the individual, which may be exclusive to certain entities. These individual-level scores can be updated and periodically uploaded to an entity (e.g., a reporting agency) that can utilize existing infrastructure to quickly provide these scores to any requesters.

A method for providing performance assessment of terminal devices is provided. A user initiates, by way of a service application that runs on a user device of the user, a first request for obtaining risk scores or connectivity scores of the terminal devices. The first request may include terminal identifiers of specific terminal devices or information pertaining to a specific geographical area. The user device communicates the first request to a server. The server determines the risk scores or the connectivity scores based on the first request. The server transmits, to the user device, a first response that includes the risk scores or the connectivity scores. The user device displays the risk scores or the connectivity scores to the user based on the first response, thereby providing the performance assessment of the terminal devices.

This application relates to apparatus and methods for identifying fraudulent transactions. In some examples, a computing device generates a decision matrix to identify fraudulent transactions. To generate the decision matrix, the computing device may determine scores for a plurality of transactions, and may determine transaction categories for each transaction based on the scores. The computing device may also determine a number of predictable features based on applying machine learning techniques to the transactions. A risk category is then determined for the number of predictable features. The computing device generates the decision matrix based on the transaction categories and the risk categories. In some examples, the computing device applies the generated decision matrix to an ongoing purchase transaction to determine if the ongoing purchase transaction is fraudulent. In some examples, the computing device prevents completion of the purchase transaction if the purchase transaction is determined to be fraudulent.

A plurality of bank identification number (BIN) ranges are characterized according to credit risk. A list of the plurality of bank identification number (BIN) ranges characterized by credit risk is made available to a transit-specific payment network interface processor, which is coupled to a plurality of memory-constrained fare gates of a transit authority. The list is configured to be distributed to the memory-constrained fare gates of the transit authority. Advantageously, the list based on BIN ranges takes up less memory than a list based on individual account numbers or the like and can be maintained in memory at the memory-constrained fare gates for rapid decisioning.

A method for analyzing credit score scenarios includes operations of receiving user input from a primary user indicating selection of a future credit score, conducting analytics on a current credit score, the future credit score, account information of the primary user, and secondary variables, to generate instructions that, when implemented, modify the account information of the primary user resulting in modification of the current credit score to the future credit score, and responsive to receiving additional input from the primary user, initiating implementation of one or more of the instructions at least through transmission of a first instruction to destination. The secondary variables may be a result of analyses of anonymized data of second users, and in some instances, indicate that alteration of a first variable within the account information of the primary user will have a greater impact in modifying the current credit score to the future credit score.