Target profiling in predictive modeling

• US 7,725,300 B2
• Filed: 04/16/2004
• Issued: 05/25/2010
• Est. Priority Date: 04/16/2004
• Status: Active Grant

• Current Assignee: Longhorn Automotive Group LLC (Alpha Alpha Intellectual Partners LLC)
• Original Assignee: Fortelligent, Inc.
• Inventors: Mansfield, Richard, Pinto, Stephen K., Hirshberg, Jay C.
• Primary Examiner(s): FREJD, RUSSELL WARREN
• Application Number:
  US10/826,453
• Publication Number:
  US 20050234760A1
• Time in Patent Office: 2,230 Days
• Field of Search: 703/2, 703/22, 706/21, 705/7, 705/10, 705/14, 382/100, 707/3, 707/5, 702/181
• US Class Current: 703/2
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0201 Market modelling; Market analysis; Collecting market data

Abstract

Models are generated using a variety of tools and features of a model generation platform. For example, in connection with a project in which a user generates a predictive model based on historical data about a system being modeled, the user is provided through a graphical user interface a structured sequence of model generation activities to be followed, the sequence including dimension reduction, model generation, model process validation, and model re-generation.

In connection with a process in which a user generates a collection of predictive models or an aggregate predictive model based on historical data about a system being modeled, profiles of aggregate targets are generated based on key contributory variables.

Claims

1. A machine-based method comprising:

for a process in which a user generates a set of predictor attributes based on historical data about a customer relationship system being modeled,

enabling the user to perform a first set of transformations on the predictor attributes of the data,

based on impact of the first set of transformations on a predictive power of the predictor attributes, enabling the user to determine whether to apply a second set of transformations to the predictor attributes to alter the impact on the predictive power,

automatically ranking performance of the predictor attributes as transformed by at least one of the first set and second set of transformations, and

using results of the ranking of the performance of the transformed predictor attributes for marketing communications to be made to customers, the customers being subject to the customer relationship system being modeled.

2. The machine-based method of claim 1 also including activating a display profile key to display the results of the ranking of the performance of the transformed predictor attributes.

3. The machine-based method of claim 1 in which the ranking of the performance of the transformed predictor attributes comprises disregarding interactions among the attributes.

4. The machine-based method of claim 1 in which the ranking is done using univariate regression analysis.

5. The machine-based method of claim 1 in which the marketing communications are presented on portable media.

6. The machine-based method of claim 1 in which the resulting marketing communications are presented through an internet protocol network.

7. A machine-based method comprising:

for a project in which a user generates a final predictive model based on a series of predictive models, the final predictive model being associated with a customer relationship system,

for the final predictive model, automatically grouping customers into segments based on identified distinguishing characteristics of the customers in the customer relationship system,

generating predictor variables for each segment,

enabling the user to apply transformations to the predictor variables until significant interactions among the generated predictor variables are taken into account, each transformed predictor variable being associated with at least one of the series of predictive models,

generating the final predictive model based on at least some of the predictive models of the series that are associated with one or more of the transformed predictor variables, and

using the final predictive model for marketing communications to be made to customers who are subject to the customer relationship system being modeled.

8. The machine-based method of claim 7 in which the distinguishing characteristics of the customers are identified by ranking a set of predictor attributes with respect to at least one target attribute.

9. The machine-based method of claim 7 in which the user is provided a graphical display of a potency of the predictor attributes in the identifying of the segments of customers.

10. The machine-based method of claim 7 in which the user is enabled to determine attributes associated with data that are associated with propensities of the customers, including enabling the user to prepare the data, transform the attributes associated with the data, and determine optimal attributes for the marketing communications.

11. The machine-based method of claim 7 in which the marketing communications are presented on portable media.

12. The machine-based method of claim 7 in which the resulting marketing communications are presented through an internet protocol network.