# US 6,658,393 B1 · Financial risk prediction systems and methods therefor

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 6,658,393 B1
- Document ID: US6658393B1
- Status: Granted
- Kind: B1
- Application number: 09/397,160
- Filed: 1999-09-15
- Granted: 2003-12-02
- Earliest claimed: 1997-05-27
- Assignees: Visa Internation Service Association
- Inventors: Catherine A. Basch; Belva J. Bruesewitz; Kevin Siegel; Patrick Faith
- Web page: https://patentgrep.com/patent/US6658393B1

## CPC

- G07F 7/1008: {Active credit-cards provided with means to personalise their use, e.g. with PIN-introduction/comparison system}
- G06Q 20/341: {Active cards, i.e. cards including their own processing means, e.g. including an IC or chip}
- G06Q 20/40: Authorisation, e.g. identification of payer or payee, verification of customer or shop credentials; Review and approval of payers, e.g. check credit lines or negative lists
- G06Q 20/40145: {Biometric identity checks}
- G06Q 20/4016: {involving fraud or risk level assessment in transaction processing}
- G06Q 20/403: {Solvency checks}
- G06Q 20/4037: {Remote solvency checks}
- G06Q 40/00: Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q 40/03: Credit; Loans; Processing thereof
- G06Q 40/08: Insurance
- G07F 7/08: by coded identity card or credit card {or other personal identification means (without personal verification means G07F7/02)}
- G07F: COIN-FREED OR LIKE APPARATUS (coin sorting G07D3/00; coin testing G07D5/00; {handling coins or paper currencies apart from payment activated apparatus G07D; payment architectures, schemes or protocols G06Q20/00})
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR

## Abstract

A method for predicting financial risk is disclosed. The method includes receiving data inputs on a first computing system. The data inputs includes historical data associated with at least a first account issued to an account owner, and the historical data includes historical transaction information for the first account. The method also includes generating a predictive model based on at least the historical data, receiving a current transaction authorization request associated with the first account on the first computing system, and generating a risk score by applying the predictive model to data associated with the current transaction authorization request. The current transaction authorization request is denied when the risk score indicates an unacceptable level of risk. In one embodiment, the data inputs further include performance data that is at least partially indicative of past fraudulent activities associated with the first account and at least one other account held by the account owner.

## Claims

### Claim 1

A computer-implemented method for predicting financial risk, the computer-implemented method comprising:
receiving data inputs on a first computing system, the data inputs are received from at least one second computing system that is separate from the first computing system, the data inputs including historical data associated with at least a first account issued to an account owner, the historical data including historical transaction information for the, first account;
generating a predictive model, the predictive model being based on at least the historical data;
receiving a current transaction authorization request associated with the first account, the current transaction authorization request being received on the first computing system;
generating a risk score, the risk score is generated by applying the predictive model to data associated with the current transaction authorization request; and
causing the current transaction authorization request to be denied when the risk score indicates an unacceptable level of risk;
wherein the data inputs further include performance data, the performance data being at least partially indicative of past fraudulent activities associated with the first account and at least one other account held by the account owner.

### Claim 2 (depends on 1)

A computer-implemented method as recited in claim 1 wherein the data inputs further include performance data, the performance data being at least partially indicative of past fraudulent activities associated with the first account.

### Claim 3 (depends on 1)

A computer-implemented method as recited in claim 1 wherein the current transaction authorization request is received from a financial authorization network.

### Claim 4 (depends on 1)

A computer-implemented method as recited in claim 1 wherein generating the predictive model includes using the historical data and the performance data.

### Claim 5 (depends on 1)

A computer-implemented method as recited in claim 1 further including:
determining when the risk score indicates an unacceptable level of risk; and
causing the current transaction authorization request to be accepted when the risk score does not indicate an unacceptable level of risk.

### Claim 6 (depends on 1)

A computer-implemented method as recited in claim 1 wherein the historical data further includes account information associated with the first account.

### Claim 7

An apparatus for predicting financial risk, the apparatus including:
a first computing unit, the first central processing unit being arranged to receive historical data regarding at least a first account issued to an account owner, the historical data including at least a transaction history for the first account;
at least a second computing unit, the second computing unit being arranged to transmit historical data to the first computing unit;
a model generator, the model generator being arranged to generate a predictive model for financial risk using at least the historical data, wherein the model generator is in communication with the first computing unit;
a receiver, the receiver being arranged to receive a current transaction authorization request associated with the first account from a financial authorization network, the receiver is coupled to the first computing unit; and
a scorer, the scorer being arranged to apply the predictive model to data contained in the current transaction authorization request to generate a risk score, the scorer is further arranged to cause the current transaction authorization request to be declined when the risk score indicates an unacceptable level of risk;
the data inputs further include performance data, the performance data being at least partially indicative of past fraudulent activities associated with the first account and at least one other account held by the account owner.

### Claim 8 (depends on 7)

An apparatus according to claim 7 wherein the model generator is further arranged to generate the predictive model using the historical data and the performance data.

### Claim 9 (depends on 7)

An apparatus according to claim 7 wherein the data inputs further include performance data the performance data being at least partially indicative of past fraudulent activities associated with the first account.

### Claim 10 (depends on 8)

An apparatus according to claim 8 wherein the historical data further includes account information associated with the first account.

### Claim 11

A computer-implemented method for predicting financial risk, the computer-implemented method comprising:
receiving first transaction data pertaining to a first transaction performed on a first credit account, said first credit account representing a credit account issued to a given account holder by a first account issuer;
receiving second transaction data pertaining to a second transaction performed on a second credit account, said second credit account being different from said first credit account, said second credit account representing a credit account issued to said given account holder by a second account issuer, said second account issuer being different from said first account issuer;
scoring said first transaction data and said second transaction data, said scoring being based on a preexisting model, to form a combined score for said given account holder;
determining when the combined score is below a predefined financial risk threshold; and
transmitting said combined score to at least one of said first account issuer and said second account issuer when it is determined that said combined score is below the predefined financial risk threshold.

### Claim 12 (depends on 11)

The computer implemented method of claim 11 further comprising:
determining when said combined score reflects a high level of financial risk;
transmitting said combined score to a third account issuer, said third account issuer being different from both said first account issuer and said second account issuer when said combined score reflects the high level of financial risk, said third account issuer representing an issuer of credit to said given account holder.

### Claim 13

A computer program product for predicting financial risk, the computer program product comprising:
computer code arranged to receive data inputs on a first computing system, wherein the data inputs are received from at least one computing system that is separate from the first computing system, the data inputs including historical data associated with at least a first account issued to an account owner, the historical data including historical transaction information for the first account;
computer code arranged to generate a predictive model based on at least the historical data;
computer code arranged to receive a current transaction authorization request associated with the first account, the current transaction authorization request being received on the first computing system from a financial authorization network;
computer code arranged to generate a risk score, the computer code arranged to generate a risk score includes computer code for applying the predictive model to data associated with the current transaction authorization request;
computer code arranged to cause the current transaction authorization request to be denied when the risk score indicates an unacceptable level of risk; and
a computer readable medium that stores the computer codes wherein the data inputs further include performance data, the performance data being at least partially indicative of past fraudulent activities associated with the first account and at least one other account held by the account owner.

### Claim 14 (depends on 13)

A computer program product according to claim 13 wherein the data inputs further include performance data, the performance data being at least partially indicative of past fraudulent activities associated with the first account.

### Claim 15 (depends on 13)

A computer program product according to claim 13 wherein the historical data further includes account information associated with the first account.

### Claim 16 (depends on 13)

A computer program product according to claim 13 wherein the computer code arranged to generate the predictive model includes computer arranged to use the historical data and the performance data to generate the predictive model.

### Claim 17 (depends on 13)

A computer program product according to claim 13 further including:
computer code arranged to determine when the risk score indicates an unacceptable level of risk; and
computer code arranged to cause the current transaction authorization request to be accepted when the risk score does not indicate an unacceptable level of risk.

### Claim 18 (depends on 11)

The computer-implemented method of claim 11 wherein a social security number of said given account holder is employed as a key to identify said first credit account and said second credit account as being issued to said given account holder.

### Claim 19 (depends on 11)

A computer-implemented method as recited in claim 11 wherein at least one of the first transaction data and the second transaction data includes current transaction data.

### Claim 20

A computer-implemented method for assessing a level of financial risk, the computer-implemented method comprising:
receiving at least one of historical data regarding at least a first account associated with an account owner and performance data regarding past fraudulent activities associated with the account owner, the historical data includes a transaction history for the first account;
generating a predictive model based at least on the at least one of the historical data and the performance data, the predictive model being arranged to predict a level of financial risk associated with the account owner,
receiving an authorization request for a current transaction involving the first account, and
generating a score using the predictive model, that includes applying the predictive model to information included in the authorization request;
further including receiving historical data regarding a second account associated with the account owner, the historical data regarding the second account includes a transaction history for the second account.

### Claim 21 (depends on 20)

A computer-implemented method as recited in claim 20 further including:
determining when the score indicates an acceptable level of financial risk;
causing the authorization request to be accepted when the score indicates the acceptable level of financial risk; and
causing the authorization request to be denied when the score does not indicate the acceptable level of financial risk.

### Claim 22 (depends on 21)

A computer-implemented method as recited in claim 21 when the authorization request is received in substantially real-time from a financial authorization network.

### Claim 23 (depends on 20)

A computer-implemented method as recited in claim 20 wherein the predictive model is further generated using the transaction history for the second account.

### Claim 24

An apparatus for assessing a level of financial risk, the apparatus comprising:
a processing unit that is arranged to receive at least one of historical data regarding at least a first account associated with an account owner and performance data regarding past fraudulent activities associated with the account owner, the historical data including a transaction history for the first account, the processing unit is further arranged to receive an authorization request for a current transaction involving the first account;
a modeler arranged to generate a predictive model based at least on the at least one of the historical data and the performance data, the predictive model being arranged to predict a level of financial risk associated with the account owner, the modeler is associated with the processing unit; and
a scorer arranged to generate a score using the predictive model, the scorer is further arranged to apply the predictive model to information included in the authorization request;
wherein the processing unit is further arranged to receive historical data regarding a second account associated with the account owner, the historical data regarding the second account includes a transaction history for the second account.

### Claim 25 (depends on 24)

An apparatus according to claim 24 wherein the scorer is further arranged to:
determine when the score indicates an acceptable level of financial risk;
cause the authorization request to be accepted when the score indicates the acceptable level of financial risk; and
cause the authorization request to be denied when the score does not indicate the acceptable level of financial risk.

### Claim 26 (depends on 24)

An apparatus according to claim 24 wherein the authorization request is received in substantially real-time from a financial authorization network.

### Claim 27 (depends on 24)

An apparatus according to claim 24 wherein the modeler is further arranged to generate the predictive model using the transaction history for the second account.

### Claim 28

A computer program product for assessing a level of financial risk, the computer program product comprising:
computer code arranged to receive at least one of historical data regarding at least a first account associated with an account owner and performance data regarding past fraudulent activities associated with the account owner, said the historical data includes a transaction history for the first account;
computer code arranged to generate a predictive model based at least on the at least one of the historical data and the performance data, the predictive model being arranged to predict a level of financial risk associated with the account owner;
computer code arranged to receive an authorization request for a current transaction involving the first account;
computer code arranged to generate a score using the predictive model, said generating the score includes applying the predictive model to information included in the authorization request; and
a computer readable medium that stores the computer codes;
further including computer code arranged to receive historical data regarding a second account associated with the account owner, said the historical data regarding the second account includes a transaction history for the second account.

### Claim 29 (depends on 28)

A computer program product according to claim 28 wherein the computer code arranged to generate the predictive model includes computer code arranged to generate the predictive model using the transaction history for the second account.

### Claim 30 (depends on 28)

A computer program product according to claim 28 further including:
computer code arranged to determine when the score indicates an acceptable level of financial risk;
computer code arranged to cause the authorization request to be accepted when the score indicates the acceptable level of financial risk; and
computer code arranged to cause the authorization request to be denied when the score does not indicate the acceptable level of financial risk.

### Claim 31 (depends on 28)

A computer program product according to claim 28 wherein the authorization request is received in substantially real-time from a financial authorization network.

### Claim 32

A computer-implemented method for predicting financial risk comprising:
receiving transaction data pertaining to at least one transaction for at least one credit account;
generating a score by applying the transaction data to a preexisting model, the preexisting model being indicative of a pattern associated with preexisting transaction data, said pattern being arranged to include events that impact the financial risk wherein said/transaction data and said pre-existing transaction data are of the same class;
determining when said score is below a predefined financial risk; and then
transmitting said score to an account issuer of said credit account when it is determined that said score is below the predefined financial risk.

### Claim 33 (depends on 32)

A method as recited in claim 32 wherein said transaction data pertains to transactions from multiple credit accounts.

### Claim 34 (depends on 32)

A computer-implemented method as recited in claim 32 wherein the transaction data and the preexisting transaction data pertain to scoreable transactions and includes account/account holder-level data pertaining to an account holder of the credit account, the credit account being one of a plurality of credit accounts issued to the account holder.

### Claim 35 (depends on 32)

A computer-implemented method as recited in claim 32 wherein said account issuer is a first account issuer, the method further including:
transmitting said score to a second account issuer of a different credit account when said score is determined to be below said predefined financial risk, said second account issuer being different from said first account issuer.

### Claim 36 (depends on 35)

A computer-implemented method as recited in claim 35 wherein said transaction data is current transaction data.

### Claim 37 (depends on 36)

A computer-implemented method as recited in claim 36 further including generating the pre-existing model, wherein the pre-existing transaction data is non-current transaction data.

### Claim 38 (depends on 32)

A computer-implemented method as recited in claim 32 further including:
archiving the transaction data.

### Claim 39 (depends on 32)

A computer-implemented method as recited in claim 32 further including:
transmitting said score to a transaction authorization system, wherein when said score is determined to be below the predefined financial risk, the transaction authorization system is arranged to deny an authorization request.

### Claim 40 (depends on 32)

A computer-implemented method as recited in claim 32 further including:
generating a new model, the new model being arranged to replace the preexisting model, wherein the new model is generated using the transaction data and the preexisting model.

### Claim 41 (depends on 32)

A computer-implemented method as recited in claim 32 wherein said transaction data includes current transaction data and said preexisting transaction data includes historical transaction data.

### Claim 42 (depends on 32)

A computer-implemented method as recited in claim 32 wherein the account/account holder-level data includes data other than data pertaining to an exchange of credit for one of goods, services, and cash.

### Claim 43 (depends on 34)

A computer-implemented method as recited in claim 34 wherein the account/account holder-level data includes public records data.

### Claim 44 (depends on 32)

A computer-implemented method as recited in claim 32 wherein the transaction data includes financial transaction data obtained in between billing cycles for a account holder of the credit account.

### Claim 45 (depends on 32)

A computer-implemented method as recited in claim 32 wherein score is generated in between billing cycles for the account issuer.
