High resolution transaction-level fraud detection for payment cards in a potential state of fraud
AssigneeFair Isaac Corporation
InventorsScott M. Zoldi, Hila Hashemi, Todd Alan Smith
Life of this patent
Estimating the term…
Abstract
A system and method are disclosed, to distinguish fraudulent transactions from a legitimate transaction, predicated on the notion that the card is considered likely to be in state of fraud. The disclosed system and method can be activated as soon as an account has suspicious activity that causes a high score for potential fraud, but before a bank either can or needs to confirm fraud. The system or method is able to pinpoint the actual fraudulent transactions inside a window of potential fraudulent activity, using a specialized model referred to as the pinpoint model.
Claims20
A method for scoring the legitimacy of specific and subsequent transactions by a payment card that is likely in a state of fraud, the method comprising:
- receiving, by a pinpoint processor, transaction data for a transaction by a payment card associated with a fraud score indicating the payment card is likely in a state of fraud, the fraud score being generated by a fraud detection computing system in communication with the pinpoint processor, the transaction data including one or more attributes;
- accessing, by the pinpoint processor from a card profile associated with the payment card, a token table having an indexed table of n most frequent tokens associated with the payment card, a frequency table of pseudo-frequencies of the corresponding n most frequent tokens and linked with the token table by a common index, and a ranking table that provides a ranking of the tokens;
- performing, by the pinpoint processor, a look-up of the token table based on the one or more attributes in the transaction data to determine a frequency and a rank of each of the one or more attributes that correspond with a token;
- calculating, by the pinpoint processor, one or more variables based on the frequency and the rank of each of the one or more attributes;
- and
- generating, by the pinpoint processor using a pinpoint model, a score that represents a likelihood of the transaction being legitimate.
A system for scoring the legitimacy of specific and subsequent transactions by a payment card that is likely in a state of fraud, the system comprising:
- a database storing a card profile associated with the payment card, the card profile including a token table having an indexed table of n most frequent tokens associated with the payment card, a frequency table of pseudo-frequencies of the corresponding n most frequent tokens and linked with the token table by a common index, and a ranking table that provides a ranking of the tokens;
- and
- a pinpoint processor connected with a fraud detection computing system for receiving transaction data for a transaction by a payment card associated with a fraud score indicating the payment card is likely in a state of fraud, the fraud score being generated by the fraud detection computing system, the transaction data including one or more attributes, the pinpoint processor further being configured to access the card profile from the database, perform a look-up of the token table based on the one or more attributes in the transaction data to determine a frequency and a rank of each of the one or more attributes that correspond with a token, calculate one or more variables based on the frequency and the rank of each of the one or more attributes, and generate, using a pinpoint model, a score that represents a likelihood of the transaction being legitimate.
A system for scoring the legitimacy of specific and subsequent transactions by a payment card that is likely in a state of fraud, the system comprising:
- at least one hardware data processor;
- and
- at least one memory storing instructions which, when executed by the at least one data processor, result in operations comprising:
- receive transaction data for a transaction by a payment card associated with a fraud score indicating the payment card is likely in a state of fraud, the fraud score being generated by a fraud detection computing system in communication with the pinpoint processor, the transaction data including one or more attributes;
- access, from a card profile associated with the payment card, a token table having an indexed table of n most frequent tokens associated with the payment card, a frequency table of pseudo-frequencies of the corresponding n most frequent tokens and linked with the token table by a common index, and a ranking table that provides a ranking of the tokens;
- execute a look-up of the token table based on the one or more attributes in the transaction data to determine a frequency and a rank of each of the one or more attributes that correspond with a token;
- calculate one or more variables based on the frequency and the rank of each of the one or more attributes;
- generate, using a pinpoint model, a score that represents a likelihood of the transaction being legitimate;
- and
- update the token table, the frequency table, and the ranking table based on the transaction data for access by the pinpoint processor for the subsequent transactions.
Dependencies are inferred from the claim text.