Geometrization For Pattern Recognition Data Analysis, Data Merging And Multiple Criteria Decision Making

• US 20110093482A1
• Filed: 12/28/2010
• Published: 04/21/2011
• Est. Priority Date: 07/30/2002
• Status: Active Grant

• Current Assignee: Abel Wolman
• Original Assignee: Abel Wolman
• Inventors: Wolman, Abel
• Granted Patent: US 8,055,677 B2
• Time in Patent Office: Days
• US Class Current: 707/756
• CPC Class Codes: G06F 18/2137 based on criteria of topology preservation, e.g. multidimensional scaling or self-organising maps
• CPC Class Codes: Y10S 707/968 Partitioning
• CPC Class Codes: Y10S 707/99943 Generating database or data structure, e.g. via user interface

Abstract

An analyzer/classifier/synthesizer/prioritizing tool for data comprises use of an admissible geometrization process with data transformed and partitioned by an input process into one or more input matrices and one or more partition classes and one or more scale groups. The data to be analyzed/classified/synthesized/prioritized is processed by an admissible geometrization technique such as 2-partition modified individual differences multidimensional scaling (2p-IDMDS) to produce at least a measure of geometric fit. Using the measure of geometric fit and possibly other 2p-IDMDS output, a back end process analyzes, synthesizes, classifies, and prioritizes data through patterns, structure, and relations within the data.

Claims

1. A computer system for data scaling comprising:

(a) a memory which is able to store data,

(b) a processor which is operatively connected to the memory which will;

(1) compute two partitions of the stored data, and

(2) compute admissible geometrization on the two partitions of the stored data, and

(3) compute admissibly transformed data so that said memory will store the transformed data, and

(c) a display to display the transformed data to a user to interpret the stored transformed data as transformed data.

2. The computer system for data scaling of claim 1 wherein the memory stores data selected from the group consisting of qualitative data, quantitative data, and intermixed qualitative and quantitative data.

3. The computer system for data scaling of claim 1 wherein in step (b2) the processor computes 2-partition individual differences multidimensional scaling on the two partitions of the stored data.

4. The computer system for data scaling of claim 1 wherein the memory stores data selected from the group consisting of:

classified data for a predetermined characteristic among a plurality of input domains, measured data for a predetermined characteristic among a plurality of input domains, survey data for a predetermined characteristic among a plurality of input domains, scoring data for a predetermined characteristic among a plurality of input domains, rating data for a predetermined characteristic among a plurality of input domains, ranking data for a predetermined characteristic among a plurality of input domains, preference data for a predetermined characteristic among a plurality of input domains, voting data for a predetermined characteristic among a plurality of input domains, and grading data for a predetermined characteristic among a plurality of input domains.

5. A data mining method comprising:

(a) using a computer processor, receiving data;

(b) forming two partitions of the received data;

(c) applying admissible geometrization to the partitioned data to produce a plurality of deformation matrices;

(d) analyzing the plurality of deformation matrices;

(e) interpreting the analyzed plurality of deformation matrices as mined data.

6. The data mining method of claim 5 wherein the received data is heterogeneous data.

7. The data mining method of claim 6 wherein the heterogeneous data is selected from the group consisting of qualitative data, quantitative data, and intermixed qualitative and quantitative data.

8. The data mining method of claim 5 wherein applying admissible geometrization to the partitioned data comprises applying 2-partition individual differences multidimensional scaling to the partitioned data.

9. The data mining method of claim 5 wherein analyzing the plurality of deformation matrices frirther comprises using clustering analysis to analyze the plurality of deformation matrices.

10. The data mining method of claim 9 wherein the clustering analysis comprises a multiresponse permutation procedure.

11. The data mining method of claim 5 wherein the received data includes unknown data values.

12. The data mining method of claim 5 wherein the analysis of the plurality of deformation matrices comprises predicting values for unknown data values.

13. The data mining method of claim 5 wherein the received data is intermixed qualitative and quantitative data.

14. Computer executable software program code stored on a computer readable medium which, when executed on a computer processor, performs the method of claim 5.

15. A computer system for data mining comprising:

(a) a memory which is able to store data,

(b) a processor which is operatively connected to the memory which will;

(1) compute two partitions of the stored data, and

(2) compute admissible geometrization of the partitioned data, and

(3) compute a plurality of deformation matrices, and

(4) compute an analysis of the plurality of deformation matrices so that said memory will store the analysis of the deformation matrices, and

(c) a display to display the analysis of the defoimation matrices to a user to interpret the stored analysis of the deformation matrices as mined data.

16. The computer system for data mining of claim 15 wherein the memory stores data that is heterogeneous data.

17. The computer system for data mining of claim 16 wherein the heterogeneous data is selected from the group consisting of qualitative data, quantitative data, and intermixed qualitative and quantitative data.

18. The computer system for data mining of claim 15 wherein the memory stores data including unknown values.

19. The computer system for data mining of claim 15 wherein in step (b4) the processor predicts values for unknown data values.

20. The computer system for data mining of claim 15 wherein in step (b2) the processor computes 2-partition individual differences multidimensional scaling on the partitioned data.