Data processing: financial – business practice – management – or co – Automated electrical financial or business practice or... – Finance
Reexamination Certificate
1994-05-05
2003-01-28
Trammell, James P. (Department: 3621)
Data processing: financial, business practice, management, or co
Automated electrical financial or business practice or...
Finance
Reexamination Certificate
active
06513018
ABSTRACT:
TECHNICAL FIELD
This invention relates generally to methods and apparatus for generating a score predictive of future performance based upon historic performance information and, more particularly, to methods and apparatus for enhancing the predictive value of such a score for a selected receiver of future performance from a selected performer by using tailored characteristics from a generic historical database including information on the performer's past performance for receivers in addition to the selected one.
BACKGROUND OF THE INVENTION
Empirical techniques that use statistical methods to predict the probability of desired performance results have heretofore been used in credit scoring, for example, to predict the probability of repayment by a credit applicant, as well as to predict other desired performance from a preselected one from an entire population of performers. Payment projection scores have been used for a number of years to assist credit issuers in prioritizing collection efforts on delinquent accounts based on internal account billing information.
Typical credit scoring systems evaluate certain historical financial and non-financial characteristics of a sample population of applicants in computer-usable form to derive a score predictive of future payment performance by a new unknown applicant who may not have been in that sample. Such scoring assumes that future performance within the population of all applicants will be consistent with past performance of the sample and identifies specific characteristics associated with past performance which collectively are most predictive of desired future performance results, such as repayment of a bank loan or credit card debt. Credit scoring, thus, assesses the credit quality of an individual about whom a decision is being made concerning the extension or use of credit. A scorecard is used to calculate the numeric value, or score, for each account considered for a credit decision.
Heretofore, developers of scoring systems have determined by discriminant analysis and other statistical techniques what combination of characteristics are positively or negatively correlated with desired performance results and have classified these predictive characteristics into attribute ranges, each having weighted score values. These attribute score values are accumulated for a particular performer, such as a credit applicant, for a set of identified characteristics to develop a total score predictive of the desired performance result.
The development and use of credit scoring, for example, has been described in a number of publications including the article entitled “Credit Scoring and the Equal Credit Opportunity Act” by David C. Hsia,
The Hastings Law Journal
, Volume 30, November 1978, No. 2, pages 382-404, and other publications.
SUMMARY OF THE INVENTION
In accord with the illustrative embodiment of this invention, the predictive value of a performance score may be enhanced by using tailored characteristics from a generic database that includes historic performance information about a selected performer relative to a number of receivers of similar performance from the performer.
The invention is particularly useful in connection with methods and apparatus for generating a collection score predictive of the likely repayment dollar amount from a selected debtor on a delinquent credit account for a selected credit issuer which uses billing and collection information for that debtor in credit bureau databases. The methods and apparatus identify the historically most predictive characteristics in a selected credit issuer database and matched generic collection information in a credit bureau database and develop a scaled score for the attributes for each relevant claracteristic. Then for a particular delinquent debtor, the delinquency and amount owed on a selected credit lender account is compared to that same information for all other accounts in the debtor's credit record to produce a score which rank orders all delinquent credit accounts according to the dollar amount likely to be repaid within a particular future time interval.
Thus, the described embodiment uses a credit bureau-based payment projection score developed from credit bureau information to predict the likely repayment dollar amount on a particular delinquent account. It effectively models a continuous performance variable of the expected dollar amount of repayment, the higher the better as the desired performance result. The amount owed on a particular credit grantor's account compared with the amount owed on delinquent accounts of all other credit grantors reflects the degree of collection competition and characteristics based on such a comparison have been found to enhance the predictive value of a collection score. So also have characteristics based on the relative level of the delinquent accounts.
One object of the invention, therefore, is to identify and use tailored characteristics as part of a generic model to fine tune the predictive capability of a score for each particular performance receiver which may differ among such receivers in the generic database.
One other object of the invention is to identify and utilize characteristics in a credit bureau database that collectively focus on predicting a collection performance as to dollar amount, rather than the odds of repayment.
Another object of the invention is by appropriate score to rank order a continuous performance variable such as the likely dollar amount collected, for multiple credit issuers in a credit bureau database.
Still other objects of the invention are a collection score that rank orders the likely repayment of multiple delinquent accounts for a credit issuer to enable collection managers to identify accounts which are more likely not to recover so as to dispose of them at an early date as well as to arrange collection queues so that accounts likely to pay a large amount are worked first before other accounts likely to pay a lesser amount.
REFERENCES:
patent: 3316395 (1967-04-01), Lavin
patent: 5239462 (1993-08-01), Jones et al.
patent: 5262941 (1993-11-01), Saladin et al.
patent: 5274547 (1993-12-01), Zaffel et al.
patent: 404247593 (1992-09-01), None
Journal of Retail Banking, vol. 8, 1986, Overstreet et al., “Manegorial Control in Credit”, pp. 79-86.*
Business Credit, vol. 90, 1988, Connors, “Credit Scoring”, pp. 51-53.*
Data Sources, 1stEdition, 1990, vol. 2, pp. J-199 to J-203.*
“Recovery Scoring with a Twist”, John Stewart, Credit Card Management, p. 12, Sep. 1992.*
“Reward Scores Credit Card Accounts for Increased Recovery”, Credit Risk Management Report, vol. 3, No. 13, Jun. 21, 1993.*
“Card profits by the Numbers”, Peter Brennan, Credit Card Management, p. 48, Sep. 1993.*
“Scoring for Profitability”, Peter Brennan, Bank Technology News, p. 16, Sep. 1993.*
Recession Changes Customer Payment Patterns Credit Risk Management Report, V2, n18, Aug. 31, 1992.*
“Recession Changes Customer Payment Patterns”, Credit Risk Management Report, v2 n18, Aug. 31, 1992.*
Robins, Gary, “Scoring: Retailers Are Finding New Uses for the Technique”, Stores, v74 n3, p. 45, Mar. 1992.*
Punch, Linda, “A Sobering Year for Collectors”, Credit Card Management, v5 n3, pp. 52-56, Jun. 1992.*
Radding et al, “Credit Scoring's New Frontier Card Marketers Know the Score”, Bank Management, v68 n9 pp. 55-64, Sep. 1992.*
Morrall, K. “Dear Deadbeat: Collections & Recovery”, Credit Card Management, v5 n8 pp. 18-21, Nov. 1992.*
“Study Says Behavior Scores Do Predict Bankruptcies”, Credit Risk Management Report, v2 n15, Jul. 20, 1992.*
“Expanded Scoring May Predict More About Account Performance”, Credit Risk Management Report, v1 n19, Sep. 16, 1991.*
“Predicting Bankruptcy Risk: Software Taps 3 Credit Bureaus”, American Banker, v158 n190, p. 12, Oct. 4, 1993.*
Fair, Isaac, “Experian/Fair, Isaac Risk Model”, Fair, Isaac and Company, Incorporated, 1998, pp. 1-80.
Elisca Pierre E.
Fair Isaac and Company, Inc.
Trammell James P.
LandOfFree
Method and apparatus for scoring the likelihood of a desired... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for scoring the likelihood of a desired..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for scoring the likelihood of a desired... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3041376