Adaptive fuzzy feature mapping

• US 6,249,779 B1
• Filed: 05/13/1998
• Issued: 06/19/2001
• Est. Priority Date: 05/13/1998
• Status: Expired due to Term

• Current Assignee: Hanger Solutions, LLC (IP Investments Group LLC)
• Original Assignee: Intelligenics LLC
• Inventors: Hitt, Ben A.
• Primary Examiner(s): Davis, George B.
• Application Number:
  US09/078,453
• Time in Patent Office: 1,133 Days
• Field of Search: 706/1, 706/900, 706/2, 382/209, 707/6, 707/5
• US Class Current: 706/1
• CPC Class Codes: G06N 7/02 using fuzzy logic computing arrangements based on biological models G06N3/00; computing arrangements using knowledge-based models G06N5/00
• CPC Class Codes: Y10S 706/90 Fuzzy logic
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

An adaptive fuzzy feature mapping (AFFM) technique provides a method for identifying and matching a new data pattern against a set of known data patterns using a combination of distance measurements and fuzzy logic functions. Known data patterns are stored as organized nodes in a pattern map wherein each organized node is defined by one or more attribute coefficients. As distance measurement is computed between a new data pattern and each organized node of the pattern map using distance measurement wherein the organized node having the smallest distance measurement to the new data pattern receives the highest ranking. Traversing the organized nodes according to the ranking, the new data pattern is compared to each organized node using fuzzy logic functions. If the new data pattern matches an organized node based on an acceptable degree of fuzziness, the attribute coefficients of the organized node are updated to reflect those coefficients of the new data pattern. If the new data pattern does not match any of the organized nodes in the pattern map, a new organized node is created in the pattern map representing the attribute coefficients of the new data pattern.

Claims

1. A method for pattern matching a new data pattern using an adaptive fuzzy feature mapping having a pattern map storing one or more organized nodes, wherein each organized node represents a known data pattern defined by one or more attribute coefficients, comprising the steps of:

a. computing a distance measurement between the new data pattern and each organized node of the pattern map;

b. ranking each organized node of the pattern map according to said distance measurement computed in step (1) for said organized node, thereby creating an ordered pattern map;

c. traversing the organized nodes of said ordered pattern map according to said ranking of step (b) by selecting a current organized node; and

d. determining whether the new data pattern matches said current organized node of said ordered pattern map by using fuzzy logic techniques and an acceptable degree of fuzziness.

2. The method for pattern matching according to claim 1, further comprising the step of:

e. adjusting the attribute coefficients of said current organized node if it is determined in step (d) that the new data pattern matches said current organized node.

3. The method for pattern matching according to claim 1, further comprising the step of:

e. creating a new organized node in the pattern map, wherein said new organized node represents the new data pattern, if it is determined in step (d) that the new data pattern does not match any organized node in the pattern map.

4. The method for pattern matching according to claim 1, wherein said step (b) ranks the organized nodes such that an organized node having a shortest distance to the new data pattern receives a highest rank, an organized node having a longest distance to the new data pattern receives a lowest rank, and the remaining organized nodes are ranked in between said highest rank and said lowest rank according to the respective said distance measurement of each organized node.

5. The method for pattern matching according to claim 4, wherein said step (c) traverses said ordered pattern map starting with said organized node having said highest rank and continues through said ordered pattern map according to said ranking and ending with said organized node having said lowest rank.

6. The method for pattern matching according to claim 1, wherein step (a) computes said distance measurement using Feature Mapping.

7. The method for pattern matching according to claim 1, wherein step (d) determines whether the new data pattern matches said current organized node using Fuzzy ART.

8. The method for pattern matching according to claim 1, further comprising the step of:

(e) inputting said acceptable degree of fussiness by an operator for determining whether the new data pattern matches said current organized node.

9. The method for pattern matching of claim 1, wherein said acceptable degree of fuzziness for determining whether the new data pattern matches said organized node using is a fixed set of one or more values.

10. The method for pattern matching according to claim 1, further comprising the step of:

e. inputting the new data pattern.

11. The method for pattern matching according to claim 1, further comprising the steps of:

e. receiving a data stream;

f. identifying a new data pattern in said data stream; and

g. inputting the new data pattern.

12. A method for pattern matching a new data pattern using an adaptive fuzzy feature mapping having a pattern map storing one or more organized nodes, wherein each organized node represents a known data pattern defined by one or more attribute coefficients, comprising the steps of:

a. comparing the new data pattern with each organized node of the pattern map using a first pattern matching technique, generating a comparison result for each organized node;

b. ranking each organized node of the pattern map according to said comparison result computed in step (a) for said organized node, thereby creating an ordered pattern map;

c. traversing the organized nodes of said ordered pattern map according to said ranking of step (b) by selecting a current organized node; and

d. determining whether the new data pattern matches said current organized node of said ordered pattern map by using a second pattern matching technique.

13. The method for pattern matching according to claim 12, wherein said first pattern matching technique comprises distance based functions and said second pattern matching technique is a fuzzy logic technique, or said first pattern matching technique is a fuzzy logic technique and said second pattern matching technique comprises distance based functions.

14. The method for pattern matching according to claim 13, wherein said distance based functions are Feature Mapping and said fuzzy logic technique if Fuzzy ART.

15. The method for pattern matching according to claim 12, further comprising the step of:

e. adjusting the attribute coefficients of said current organized node if it is determined in step (d) that the new data pattern matches said current organized node.

16. The method for pattern matching according to claim 12, further comprising the step of:

e. creating a new organized node in the pattern map, wherein said new organized node represents the new data pattern, if it is determined in step (d) that the new data pattern does not match any organized node in the pattern map.

17. A computer program product for use with a computer system, comprising:

a computer usable medium having computer readable program code means embodied in said medium for pattern matching a new data pattern using an adaptive fuzzy feature mapping having a pattern map storing one or more organized nodes, wherein each organized node represents a known data pattern defined by one or more attribute coefficients, said computer readable program code means comprising;

computing means for enabling a processor to compute a distance measurement between the new data pattern and each organized node of the pattern map;

ranking means for enabling a processor to rank each organized node of the pattern map according to said distance measurement computed by said computing means for said organized node, thereby creating an ordered pattern map;

traversing means for enabling a processor to traverse the organized nodes of said ordered pattern map according to said ranking by selecting a current organized node; and

determining means for enabling a processor to determine whether the new data pattern matches said current organized node of said ordered pattern map by using fuzzy logic techniques and an acceptable degree of fuzziness.

18. The computer program product according to claim 17, said computer readable program code means further comprising:

adjusting means for adjusting the attribute coefficients of said current organized node if it is determined by said determining means that the new data pattern matches said current organized node.

19. The computer program product according to claim 17, said computer readable program code means further comprising:

creating means for creating a new organized node in the pattern map, wherein said new organized node represents the new data pattern, if it is determined by said determining means that the new data pattern does not match any organized node in the pattern map.

20. The computer program product according to claim 17, wherein said ranking means ranks the organized nodes such that an organized node having a shortest distance to the new data pattern receives a highest rank, an organized node having a longest distance to the new data pattern receives a lowest rank, and the remaining organized nodes are ranked in between said highest rank and said lowest rank according to the respective said distance measurement of each organized node.

21. The computer program product according to claim 20, wherein said traversing means traverses said ordered pattern map starting with said organized node having said highest rank and continues through said ordered pattern map according to said ranking and ending with said organized node having said lowest rank.

22. The computer program product according to claim 17, wherein said computing means computes said distance measurement using Feature Mapping.

23. The computer program product according to claim 17, wherein said determining means determines whether the new data pattern matches said current organized node using Fuzzy ART.

24. The computer program product according to claim 17, said computer readable program code means further comprising:

inputting means for inputting the new data pattern.

25. The computer program product according to claim 17, said computer readable program code means further comprising:

receiving means for receiving a data stream,

identifying means for identifying a new data pattern in said data stream, and

inputting means for inputting the new data pattern.