Selecting a conclusion using an ordered sequence of discriminators

• US 7,908,231 B2
• Filed: 06/12/2007
• Issued: 03/15/2011
• Est. Priority Date: 06/12/2007
• Status: Active Grant

• Current Assignee: James R. Miller III
• Original Assignee: James R. Miller
• Inventors: Miller, James R.
• Primary Examiner(s): Holmes; Michael B.
• Assistant Examiner(s): CHANG, LI WU
• Application Number:
  US11/761,878
• Publication Number:
  US 20080313223A1
• Time in Patent Office: 1,372 Days
• Field of Search: 706/16
• US Class Current: 706/16
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16Z 99/00 Subject matter not provided for in other main groups of this subclass

Abstract

A method for data analysis according to various aspects of the present invention generally includes selecting a conclusion from a plurality of conclusions for one or more test data sets by generating discriminators from one or more threshold indicators associated with the conclusion, selecting a portion of the discriminators, ranking the discriminators in a sequence, and applying the sequence of discriminators to one or more test data sets to select the conclusion.

Claims

1. A method for selecting a conclusion from a plurality of conclusions, the method comprising:

generating, by a computer system, a plurality of discriminators from one or more provided threshold indicators associated with the conclusion, wherein the plurality of discriminators includes a pair of mutually exclusive and collectively exhaustive discriminators for each of the one or more provided threshold indicators;

selecting a portion of the plurality of discriminators; and

ranking the portion of the plurality of discriminators to create an ordered sequence of discriminators, wherein the ordered sequence of discriminators is for selecting the conclusion for one or more test data sets;

wherein at least one of selecting and ranking the portion of the plurality of discriminators includes;

recalculating thresholds for one or more indicators; and

generating new discriminators based on the recalculated thresholds.

2. The method of claim 1, wherein both the selecting and ranking of the portion of the plurality of discriminators is based at least partially on each discriminator'"'"'s propensity to indicate one or more of the plurality of conclusions for at least one training data set.

3. The method of claim 1, wherein generating the plurality of discriminators includes selecting a threshold for each of the one or more provided threshold indicators.

4. The method of claim 2, wherein at least one of the threshold indicators comprises a result of a data analysis procedure performed on the training data set.

5. The method of claim 2, wherein the training data set comprises a plurality of data sources.

6. The method of claim 1, further comprising selecting the one or more provided threshold Indicators.

7. The method of claim 6, wherein selecting the one or more provided threshold indicators is at least partially based on at least one of a desired sensitivity, a desired specificity, and on each provided indicator'"'"'s propensity to indicate one or more of the conclusions from the plurality of conclusions.

8. The method of claim 6, wherein selecting the one or more provided threshold indicators includes determining whether a threshold indicator includes a plurality of measurement values, the measurement values being mutually exclusive and collectively exhaustive.

9. The method of claim 8, wherein selecting the one or more provided threshold indicators further includes determining whether the plurality of measurement values are statistically associated with the one or more conclusions.

10. The method of claim 6, wherein selecting the one or more provided threshold indicators includes determining for each threshold indicator one or more probabilities that the threshold indicator is indicative of the one or more conclusions.

11. The method of claim 1, further comprising providing the one or more threshold indicators.

12. The method of claim 1, wherein the plurality of conclusions comprise a focal conclusion and one or more alternative conclusions, and wherein each of the plurality of threshold indicators includes a first range of values indicative of the focal conclusion and a second range of values indicative of the one or more alternative conclusions.

13. The method of claim 12, wherein the pair of mutually exclusive and collectively exhaustive discriminators for each threshold indicator includes a first discriminator indicative of the focal conclusion and a second discriminator indicative of the one or more alternative conclusions.

14. The method of claim 12, wherein at least one of selecting the plurality of discriminators and ranking the plurality of discriminators is based on each discriminator'"'"'s propensity to distinguish between the focal conclusion and the one or more alternative conclusions.

15. The method of claim 1, further comprising selecting one or more provided non-threshold indicators.

16. The method of claim 15, wherein selecting one or more provided non-threshold indicators includes determining for each non-threshold indicator one or more probabilities that the non-threshold indicator is indicative of the one or more conclusions.

17. The method of claim 15, wherein the plurality of discriminators includes multiple pairs of mutually exclusive and collectively exhaustive discriminators for each of the one or more provided non-threshold indicators.

18. The method of claim 15, wherein at least one of the provided non-threshold indicators comprises a result of a data analysis procedure performed on the training data set.

19. The method of claim 15, further comprising providing the one or more non-threshold indicators.

20. The method of claim 1, further comprising applying the sequence of discriminators to the one or more test data sets to select a conclusion from the plurality of conclusions for the one or more test data sets.

21. The method of claim 20, wherein the training data set comprises a plurality of data sources.

22. A method for selecting a conclusion from a plurality of conclusions using a sequence of provided discriminators, wherein the sequence of provided discriminators are selected and ranked from a plurality of discriminators into an ordered sequence, wherein at least one of selecting and ranking the portion of the plurality of discriminators includes:

recalculating thresholds for one or more indicators; and

generating new discriminators based on the recalculated thresholds, and wherein the plurality of discriminators includes a pair of mutually exclusive and collectively exhaustive discriminators generated for each of one or more threshold indicators associated with the conclusion, the method comprising;

applying the ordered sequence of provided discriminators to one or more test data sets to select the conclusion from the plurality of conclusions for the one or more test data sets.

23. A computer-readable medium having computer-executable instructions for performing a method for selecting a conclusion from a plurality of conclusions, the method comprising:

generating a plurality of discriminators from one or more provided threshold indicators associated with the conclusion, wherein the plurality of discriminators includes a pair of mutually exclusive and collectively exhaustive discriminators for each of the one or more provided threshold indicators;

selecting a portion of the plurality of discriminators; and

ranking the portion of the plurality of discriminators to create an ordered sequence of discriminators, wherein the ordered sequence of discriminators is for selecting the conclusion for one or more test data sets;

wherein at least one of selecting and ranking the portion of the plurality of discriminators includes;

recalculating thresholds for one or more indicators; and

generating new discriminators based on the recalculated thresholds.

24. A system for selecting a conclusion from a plurality of conclusions,

the system comprising:

a processor;

a memory coupled to the processor and storing instructions that, when executed by the processor, cause the processor to;

generate a plurality of discriminators from one or more provided threshold indicators associated with the conclusion, wherein the plurality of discriminators includes a pair of mutually exclusive and collectively exhaustive discriminators for each of the one or more provided threshold indicators;

select a portion of the plurality of discriminators; and

rank the portion of the plurality of discriminators to create an ordered sequence of discriminators, wherein the ordered sequence of discriminators is for selecting the conclusion for one or more test data sets;

wherein at least one of selecting and ranking the portion of the plurality of discriminators includes;

recalculating thresholds for one or more indicators; and

generating new discriminators based on the recalculated thresholds.