Statistical impact analysis computer system

• US 6,192,319 B1
• Filed: 04/24/1998
• Issued: 02/20/2001
• Est. Priority Date: 04/24/1998
• Status: Expired due to Term

• Current Assignee: Verint Americas Incorporated (Verint Systems Incorporated)
• Original Assignee: CFI Group, Inc.
• Inventors: Zhao, Jun, Simonson, Mark C., Cha, Jaesung
• Primary Examiner(s): ASSOUAD, PATRICK J
• Application Number:
  US09/066,221
• Time in Patent Office: 1,033 Days
• Field of Search: 702/81, 702/84, 702/179, 702/181, 705/1, 705/10, 707/1, 707/104
• US Class Current: 702/81
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising

Abstract

A computer-implemented apparatus and method for determining the impacts of predetermined customer characteristics associated with measured physical attributes. A manifest variable database is utilized for storing manifest variable data that is indicative of the measured physical attributes. A partial least squares determinator is connected to the manifest variable database for determining statistical weights based upon the stored manifest variable data. A weights database which is connected to the partial lease squares determinator is utilized for storing the determined weights. A latent variable determinator is connected to the weighing database for determining scores for latent variables based upon the stored manifest variables and upon the stored weights. The latent variables are indicative of the predetermined customer characteristics. Additionally, a latent variable database is connected to the latent variable score determinator for storing the determined latent variable scores. An impact determinator is connected to the latent variable database for determining impact relationships among the latent variables based upon the stored latent variable scores.

Claims

1. A computer-implemented apparatus for determining the impacts of predetermined customer characteristics associated with measured physical attributes, comprising:

a manifest variable database for storing manifest variable data that is indicative of the measured physical attributes;

a partial least squares determinator connected to said manifest variable database for determining statistical weights based upon said stored manifest variable data;

a weights database connected to said partial least squares determinator for storing said determined weights;

a latent variable determinator connected to said weighing database for determining scores for latent variables based upon said stored manifest variables and upon said stored weights, said latent variables being indicative of the predetermined customer characteristics;

a latent variable database connected to said latent variable score determinator for storing said determined latent variable scores; and

an impact determinator connected to said latent variable database for determining impact relationships among said latent variables based upon said stored latent variable scores.

2. The apparatus of claim 1 wherein said manifest variable data includes performance metric data, said impact determinator determining impact relationships of said latent variables upon said performance metric data based upon said latent variable scores and upon said performance metric data.

3. The apparatus of claim 1 further comprising:

a historical weights model database for storing superset weights related to said manifest variables;

a weight substitution determinator connected to said partial least squares determinator for selecting said superset weights based upon a predetermined criterion.

4. The apparatus of claim 3 wherein said manifest variable data are associated with a plurality of cases associated with said measured physical attributes, wherein said predetermined criteria include the numbers of cases associated with said manifest variable data.

5. The apparatus of claim 4 further comprising:

a weight average for determining the average of said selected weights, said latent variable determinator determining said latent variable scores based upon said manifest variable data and upon the average of said selected weights.

6. The apparatus of claim 5 further comprising:

a manifest variable normalizer for normalizing said manifest variable data to a predetermined scale, said latent variable determinator determining said latent variable scores based upon said normalized manifest variable data and upon the average of said normalized weights.

7. The apparatus of claim 1 further comprising:

first diagnostic rules connected to said partial least squares determinator for determining performance of said partial least squares determinator based upon a predetermined criteria.

8. The apparatus of claim 7 wherein said partial least squares determinator selects weights based upon said first diagnostic rules.

9. The apparatus of claim 1 further comprising:

second diagnostic rules connected to said latent variable determinator for determining performance of said latent variable determinator based upon predetermined criteria.

10. The apparatus of claim 9 wherein said latent variable determinator interprets latent variable scores based upon said second diagnostic rules.

11. The apparatus of claim 1 further comprising:

third diagnostic rules connected to said impact determinator to determine performance of said impact analyzer based upon predetermined criteria.

12. The apparatus of claim 11 wherein said impact determinator interprets said impact relationships based upon said third diagnostic rules.

13. The apparatus of claim 1 wherein said impact determinator uses regression analysis to determine said impact relationships.

14. The apparatus of claim 13 wherein said regression analysis includes analysis being selected from the group consisting of phantom regression analysis, constrained regression analysis, multi-variate regression analysis, and combinations thereof.

15. The apparatus of claim 1 wherein said impact determinator produces output being selected from the group consisting of direct impact output, total impact output, R-square output, adjusted R-square output, intercept analysis output, and combinations thereof.

16. The apparatus of claim 1 wherein said manifest variable data is associated with time data, said impact determinator determining the impact of the measured physical attributes in a temporal domain using said manifest variable data and said associated time data.

17. The apparatus of claim 1 further comprising:

a models database for storing a plurality of models for establishing interrelationships among said manifest variable data, said latent variable determinator determining scores for said latent variables using said model specifications.

18. The apparatus of claim 17 further comprising:

diagnostic rules for comparing one of said plurality of models with another model in said plurality of models to interpret the degree of fit between said models.