Data processing: artificial intelligence – Neural network – Learning task
Reexamination Certificate
2000-06-19
2003-04-29
Follansbee, John A. (Department: 2121)
Data processing: artificial intelligence
Neural network
Learning task
C705S038000
Reexamination Certificate
active
06556979
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Technical Field
The present invention is related to consumer credit. More particularly the present invention is a system, method and instruction for identifying which potential customers are credit revolvers for card issuers.
2. Description of Related Art
Strong competition in credit card industry has driven banks and credit card issuers to compete for a few profitable customers. Credit card issuers make their profit out of approximately 20% of total portfolio customers; most of them are revolvers. The major portion of profits for credit card issuers is from higher interest rates charged to credit revolvers. Credit revolvers or revolvers are customers who carry a balance on their credit card accounts from month to month. In general, revolvers are more profitable customers than those who do not carry a monthly balance on their credit card accounts.
Credit card utilization is the percentage of total available credit being utilized by a customer. Neither very high nor very low revolving utilization is considered to be good by credit card issuers. Revolvers with very high credit utilization are considered to be high-risk customers by most credit card issuers. Very high credit utilization may indicate that a customer is in the midst of financial difficulty. Very high credit utilization may put the credit card issuers at high risk for charge-off, or bankruptcy.
Another undesirable credit utilization pattern is an explosive growth pattern. Explosive credit utilization patterns may also lead to delinquency, charge-off, or bankruptcy because the pattern is indicative of a deleterious change in a customer's financial state. Therefore, credit card issuers must also avoid all potential customers, including revolvers, who demonstrate explosive credit utilization growth unless they have very good strategy to manage high risk customers.
Conversely, most customers with low or even zero utilization do not represent a credible risk to credit card issuers immediately, however, neither do these customers contribute to the credit card issuers' profits. Customers that maintain credit card accounts merely for the convenience of having instant credit available while shopping are known as convenient users. Normally these customers immediately pay their purchases off. Credit issuers must cover relatively high operation costs for administering customer accounts.
It would be advantageous to provide credit card issuers with the means to filter out unprofitable customers from targeting groups. It would further be advantageous to provide credit card issuers with the means to filter out high-risk customers. It would still further be advantageous to provide a system, method and instructions for objectively filtering out unprofitable customers while acquiring profitable customers.
SUMMARY OF THE INVENTION
The present invention relates to a method, system and instructions for generating a stable, time series segmentation credit utilization chart and identifying profitable revolvers from that chart. The process begins by acquiring customer time series credit files. The credit files are organized in a data mart environment that supports a query system. Time series utilization attributes are created first. Then a neural network time series segmentation algorithm is applied to the data. A resultant N×N dimension segments chart is generated for analyses. The chart may be further modified to more accurately depict profitable credit revolvers. Profitable credit revolvers are identified by having credit utilization patterns belonging to one or more of the segments in the N×N chart previously identified.
REFERENCES:
patent: 5262941 (1993-11-01), Saladin et al.
patent: 5521813 (1996-05-01), Fox et al.
patent: 5696907 (1997-12-01), Tom
patent: 5761442 (1998-06-01), Barr et al.
patent: 5870721 (1999-02-01), Norris
patent: 5933817 (1999-08-01), Hucal
patent: 6018723 (2000-01-01), Siegel et al.
Conference Title: “Proceedings of the International Conference on Artificial Neural Nets and Genetic Algorithms”, Date: Apr. 14-16, 1993, Innsbruck, Austria.
Purcell, L. Roping in risk (neural networks in banks), Journal: Bank Systems+Technology vol. 31, No. 5, p. 64-66, Pub Date: May 1994, USA ISSN, 1045-9472.
Liu Shiping
Murphy Edgar Leroy
Nanek David Wayne
Rothman Michael
Carstens Yee & Cahoon LLP
Doudnikoff Gregory M.
Follansbee John A.
Hirl Joseph H.
LandOfFree
Method and system for identifying consumer credit revolvers... 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 system for identifying consumer credit revolvers..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for identifying consumer credit revolvers... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3093109