Predictive modeling of consumer financial behavior

• US 6,430,539 B1
• Filed: 05/06/1999
• Issued: 08/06/2002
• Est. Priority Date: 05/06/1999
• Status: Active Grant

• Current Assignee: Callahan Cellular LLC (Intellectual Ventures LLC)
• Original Assignee: HNC Software Inc. (Fair Isaac Corporation)
• Inventors: Sitze, Kevin L., Lazarus, Michael A., Blume, A. U. Mattias, Brown, Kenneth B., Caid, William R., Dunning, Ted E., Peranich, Larry S., Russell, Gerald R.
• Primary Examiner(s): Choi, Kyle J.
• Assistant Examiner(s): Robertson, D
• Application Number:
  US09/306,237
• Time in Patent Office: 1,188 Days
• Field of Search: 705/10, 705/14, 705/26, 706/6
• US Class Current: 705/7.31
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0202 Market predictions or forecasting for commercial activities
• CPC Class Codes: G06Q 30/0207 Discounts or incentives, e.g. coupons or rebates
• CPC Class Codes: G06Q 30/0241 Advertisements
• CPC Class Codes: G06Q 30/0255 based on user history
• CPC Class Codes: G06Q 30/0269 based on user profile or attribute
• CPC Class Codes: G06Q 30/0601 Electronic shopping [e-shopping]

Abstract

Predictive modeling of consumer financial behavior is provided by application of consumer transaction data to predictive models associated with merchant segments. Merchant segments are derived from consumer transaction data based on co-occurrences of merchants in sequences of transactions. Merchant vectors representing specific merchants are clustered to form merchant segments in a vector space as a function of the degree to which merchants co-occur more or less frequently than expected. Each merchant segment is trained using consumer transaction data in selected past time periods to predict spending in subsequent time periods for a consumer based on previous spending by the consumer. Consumer profiles describe summary statistics of consumer spending in and across merchant segments. Analysis of consumers associated with a segment identifies selected consumers according to predicted spending in the segment or other criteria, and the targeting of promotional offers specific to the segment and its merchants.

Claims

1. A method of predicting financial behavior of consumers, comprising:

generating from transaction data for a plurality of consumers, a date ordered sequence of transactions for each consumer;

selecting for each consumer a set of the date ordered transactions to form a group of input transactions for the consumer; and

for each consumer, applying the input transactions of the consumer to each of a plurality of merchant segment predictive models, each merchant segment predictive model defining for a group of merchants a prediction function between input transactions in a past time interval and predicted spending in a subsequent time interval, to produce for each consumer a predicted spending amount in each merchant segment.

2. The method of claim 1, further comprising:

for each consumer, associating the consumer with the merchant segment for which the consumer had the highest predicted spending relative to other merchant segments.

3. The method of claim 1, further comprising:

for each merchant segment, determining a segment vector as a summary vector of merchant vectors of merchants associated with the segment; and

for each consumer, associating the consumer with the merchant segment having the greatest dot product between the segment vector of the segment and a consumer vector of the consumer.

4. The method of claim 1, further comprising:

for each merchant segment;

ranking the consumers by their predicted spending in the merchant segment;

determining for each consumer a percentile ranking in the merchant segment;

for each consumer;

determining the merchant segment in which the consumer'"'"'s percentile ranking is the highest, to uniquely associate each consumer with one merchant segment; and

for each merchant segment, determining summary transaction statistics for the consumers uniquely associated with the merchant segment.

5. The method of claim 1, further comprising:

for each merchant segment;

ranking the consumers by their predicted spending in the merchant segment;

determining for each consumer a percentile ranking in the merchant segment;

selecting as a population, the consumers having a percentile ranking in excess of predetermined percentile threshold; and

determining summary transaction statistics for selected population of consumers.

6. The method of claim 1, further comprising:

establishing for each merchant in the transaction data a merchant vector; and

updating the merchant vector of each merchant relative to the merchant vectors of other merchants according to co-occurrences of each merchant in the transaction data.

7. The method of claim 6, further comprising:

updating the merchant vector of each merchant based upon an unexpected amount deviation in a frequency of co-occurrence of the merchant with other merchants.

8. The method of claim 6, further comprising:

determining a co-occurrence frequency for each merchant with each other merchant in the transaction data;

determining for each pair of merchants, a relationship strength between the pair of merchants based on how much the determined co-occurrence frequency deviates from an expected co-occurrence frequency;

for each pair of merchant vectors, mapping the relationship strength into a vector space as a desired dot product between respective merchant vectors the merchants in the pair; and

updating each merchant vectors so that the actual dot products between each pair of merchant vectors approximates the desired dot product between the merchant vectors.

9. The method of claim 8, wherein determining for each pair of merchants, a relationship strength between the pair of merchants further comprises:

determining the relationship strength by $r_{\mathrm{ij}}=\frac{T_{\mathrm{ij}}-{\hat{T}}_{\mathrm{ij}}}{{\mathrm{\sigma <br><br>}}_{\mathrm{ij}}}\approx <br><br>\frac{T_{\mathrm{ij}}-{\hat{T}}_{\mathrm{ij}}}{\sqrt{{\hat{T}}_{\mathrm{ij}}}}$

where

r_ij is the relationship strength between merchant_i and merchant_j in a pair of merchants;

T_ij is the actual co-occurrence frequency of merchant_i and merchant_j in the transaction data; and

{circumflex over (T)}_ij is the expected co-occurrence frequency of merchant_i and merchant_j in the transaction data.

10. The method of claim 8, wherein determining for each pair of merchants, a relationship strength between the pair of merchants further comprises:

determining the relationship strength by

11. The method of claim 8, wherein determining for each pair of merchants, a relationship strength between the pair of merchants further comprises:

determining the relationship strength by $r_{\mathrm{ij}}=\mathrm{sign}<br><br>\text{\hspace{1em}<br><br>}<br><br>\left(T_{\mathrm{ij}}-{\hat{T}}_{\mathrm{ij}}\right)<br><br>\text{\hspace{1em}<br><br>}·<br><br>\text{\hspace{1em}<br><br>}<br><br>\sqrt{2<br><br>\ln <br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{\lambda <br><br>}/\sqrt{{\hat{T}}_{\mathrm{ij}}}}=\mathrm{sign}<br><br>\text{\hspace{1em}<br><br>}<br><br>\left(T_{\mathrm{ij}}-{\hat{T}}_{\mathrm{ij}}\right)<br><br>\text{\hspace{1em}<br><br>}·<br><br>\text{\hspace{1em}<br><br>}<br><br>\sqrt{2<br><br>\ln <br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{\lambda <br><br>}}<br><br>\text{\hspace{1em}<br><br>}·<br><br>\text{\hspace{1em}<br><br>}<br><br>{{\hat{T}}_{\mathrm{ij}}}^{1/4}$

where

r_ij is the relationship strength between merchant_i and merchant_j in a pair of merchants;

λ

is a log-likelihood ratio;

T_ij is the actual co-occurrence frequency of merchant_i and merchant_j in the transaction data; and

{circumflex over (T)}_ij is the expected co-occurrence frequency of merchant_i and merchant_j in the transaction data.

12. The method of claim 8, wherein updating each merchant vectors so that the actual dot products between each pair of merchant vectors approximates the desired dot product between the merchant vectors comprises a gradient descent update that updates the merchant vectors according to whether the actual dot product between them is greater or lesser than the desired dot product.

13. The method of claim 8, wherein updating each merchant vectors so that the actual dot products between each pair of merchant vectors approximates the desired dot product between the merchant vectors comprises determining for each merchant vector an error weighted average of the desired positions of the merchant vector from current position of each other merchant vector and the desired dot product between the merchant vector and each other merchant vector.

14. The method of claim 1, further comprising:

determining for each merchant name in the transaction data a merchant vector;

clustering the merchant vectors to form a plurality of merchant segments, wherein each merchant vector is associated with one and only one merchant segment;

for each merchant segment, determining from the transactions of consumers at the associated merchants of the merchant, statistical measures of consumer transactions in the segment.

15. The method of claim 1, further comprising:

selecting a plurality of consumers associated with at least one merchant segment, the selected plurality selected according to their predicted spending in the merchant segment; and

providing promotional offers to the selected plurality of consumers.

16. The method of claim 1, further comprising:

training each of the merchant segment predictive models to predict spending in a predicted time period based upon transaction statistics of the consumer'"'"'s transactions in a past time period.

17. The method of claim 16, wherein the transaction statistics comprises variables describing the recency of the consumer'"'"'s transactions in one or more merchant segments, the frequency of the consumer'"'"'s transactions in one or more merchant segments, and the amount of the consumer'"'"'s transactions in one or more merchant segments.

18. A system for predicting consumer financial behavior, comprising:

a plurality of merchant segments, each merchant segment having a set of merchants associated therewith;

a plurality of merchant segment predictive models, each model associated with one of the merchant segments for predicting spending by an individual consumer in the merchant segment in a predicted time period as a function of transaction statistics of the consumer for transactions in a prior time period; and

a data processing module that receives transaction data for a consumer, and constructs the transaction statistics for the prior time period for input into selected ones of the merchant segment predictive models.

19. A system for forming merchant segments, comprising:

a data processing module that receives consumer transaction data for a plurality of consumer accounts, and organizes the transaction data by account, and within account, sequences the transactions by time;

a data processing module that determines from the sequenced transaction data an expected frequency of co-occurrence for each merchant, and that constructs for each merchant a merchant vector as a function of unexpected frequency of co-occurrences of the merchant; and

a clustering module that clusters the merchant vectors into merchant segment by determining merchant vectors that are closely aligned with each other.