Method and apparatus for comparison of data strings

• US 5,742,706 A
• Filed: 12/07/1995
• Issued: 04/21/1998
• Est. Priority Date: 03/10/1992
• Status: Expired due to Term

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle Corporation
• Inventors: Yu, Francis H.
• Primary Examiner(s): Boudreau, Leo
• Assistant Examiner(s): Tran, Phuoc
• Application Number:
  US08/568,557
• Time in Patent Office: 866 Days
• Field of Search: 382/209, 382/182, 382/205, 382/206, 382/227, 382/229, 382/276, 382/278, 382/309, 382/310, 382/231, 364/715.04, 364/715.09, 364/715.11, 364/419.12, 395/2, 395/51, 395/54, 395/606, 395/612, 395/795, 434/167
• US Class Current: 382/229
• CPC Class Codes: G06F 16/90344 by using string matching techniques
• CPC Class Codes: G06F 2207/025 String search, i.e. pattern matching, e.g. find identical word or best match in a string
• CPC Class Codes: G06F 7/02 Comparing digital values G06F7/06, G06F7/22, G06F7/38 take precedence
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

The present invention is a method and apparatus that measures the similarity of two images. Any information that can be discretely symbolized can be transformed into an image through so-called "image projection". This process is used to define otherwise discrete entities as part of a linear space, making it possible to calculate distances among those entities. A mechanism called a cluster allows association of otherwise discrete symbols, improving the matching abilities of the invention. Initially, the sequence of symbols is normalized. Then, a projection of the normalized sequence is created. The projection may be optionally generated with a cluster that assigns weights to the neighbors of a core symbol and/or with position weights that assigns weights to each position in the normalized image. Projection matching is then performed to determine match candidates for the string of symbols.

Claims

1. In a computer system having a processing means coupled to a storage means and a display means, a computer-implemented method of comparing, in a database application, a first string of digitally represented characters with a second string of digitally represented characters, each character of said first and second string of digitally represented characters being a member of a set of digitally represented characters, said computer-implemented method comprising the steps of:

using said processing means, normalizing said first string of S known digitally represented characters to create a first normalized string of N normalized symbols and normalizing said second string of known digitally represented characters to create a second normalized string of N normalized symbols;

storing said first normalized string and said second normalized string in said storage means;

using said processing means, generating a first projection from said first normalized string and a second projection from said second normalized string;

storing said first projection and said second projection in said storage means; and

using said processing means, retrieving said first and second projections from said storage means and comparing said first projection and said second projection to determine a degree of similarity of said first and second projections.

2. The computer implemented method of claim 1 wherein said set of digitally represented characters is a set of ASCII characters.

3. The computer implemented method of claim 1 wherein said set of digitally represented characters is a set of symbols.

4. The computer implemented method of claim 1 wherein said steps of generating said first projection and said second projection further include the step of using cluster tables for generating said first projection and said second projection.

5. The computer implemented method of claim 1 wherein said steps of generating said first projection and said second projection further include the step of using position weight tables for generating said first projection and said second projection.

6. The computer implemented method of claim 1 wherein said steps of generating said first projection and said second projection further include the step of using cluster tables and position weight tables for generating said first projection and said second projection.

7. The computer implemented method of claim 1 wherein said steps of normalizing said first string and said second string further include the step of generating a medium of a symbol, M, in a normalized image by:

space="preserve" listing-type="equation">M(S.sub.i)=i*|N|/|S|

where S_i is the i-th symbol in a string S, |N| is the normalized size, and |S| is the length of string S.

8. The computer-implemented method of claim 1 wherein said step of generating said first projection and said second projection is accomplished by projecting said first string and said second string onto a first and second closure in a normal distribution by:

space="preserve" listing-type="equation">C.sub.S.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =d.sub.j

space="preserve" listing-type="equation">C.sub.S.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, and Cs_ik is the k-th item in symbol S_i '"'"'s closure.

9. The computer implemented method of claim 4 wherein said step of generating said projection of said first string and said second string using cluster tables is accomplished by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =d.sub.j *us.sub.in

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j *us.sub.in (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, Cs_ik is the k-th item in symbol S_i '"'"'s closure, and us_in is weight of symbol n in the cluster whose core is S_i.

10. The computer implemented method of claim 5 wherein said step of generating said projection of said first string and said second string using position weight tables is accomplished by

11. The computer-implemented method of claim 6 wherein said step of generating said projection of said first string and said second string using cluster tables and weight tables is accomplished by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =d.sub.j *W.sub.M(s.sbsb.i.sub.) *us.sub.i.spsb.n

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j *W.sub.M(s.sub.i.sub.) *us.sub.i.spsb.n (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, C_S.sbsb.ik is the k-th item in symbol S_i '"'"'s closure, u_s.sbsb.i_n is a cluster weight, and W_M(s.sbsb.i.sub.) is a position weight.

12. A computer apparatus for comparing, in a database application, a first string of digitally represented characters with a second string digitally represented characters, each character of said first and second string of digitally represented characters being a member of a set of digitally represented characters, said computer apparatus comprising:

a storage means;

a processing means coupled to said storage means, said processing means including;

a means for normalizing a first string of S known digitally represented characters to create a first normalized string of N normalized symbols and for normalizing a second string of known digitally represented characters to create a second normalized string of N normalized symbols;

a means for storing said first normalized string and said second normalized string;

a means for generating a first projection from said first normalized string and a second projection from said second normalized string;

a means for storing said first projection and said second projection in said storage means;

a means for retrieving said first and second projections from said storage means; and

a means for comparing said first projection and said second projection to determine a degree of similarity of said first and second projections.

13. The computer apparatus of claim 12 wherein said set of digitally represented characters is a set of ASCII characters.

14. The computer apparatus of claim 12 wherein said set of digitally represented characters is a set of symbols.

15. The computer apparatus of claim 12 wherein said means for generating said first projection and said second projection uses cluster tables.

16. The computer apparatus of claim 12 wherein said means for generating said first projection and said second projection uses position weight tables.

17. The computer apparatus of claim 12 wherein said means for generating said first projection and said second projection uses cluster tables and position weight tables.

18. The computer apparatus of claim 12 wherein said means for normalizing said first string and said second string generates a medium of a symbol, M, in a normalized image by:

space="preserve" listing-type="equation">M(S.sub.i)=i*|N|/|S|

where S_i is the i-th symbol in a string S, |N| is the normalized size, and |S| is the length of string S.

19. The computer apparatus of claim 12 wherein said means for generating said projection of said first string and said second string is accomplished by projecting said first string and said second string onto a first and a second closure in a normal distribution by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =dj

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, and C_S.sbsb.ik is the k-th item in symbol S_i '"'"'s closure.

20. The computer apparatus of claim 15 wherein said means for generating said projection of said first string and said second string uses said cluster tables by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =d.sub.j *us.sub.i.spsb.n

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j *us.sub.i.spsb.n (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, C_S.sbsb.ik is the k-th item in symbol S_i '"'"'s closure, and u_s.sbsb.i_n is a cluster weight.

21. The computer apparatus of claim 16 wherein said processing means for generating said projection of said first string and said second string uses position weight tables by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =d.sub.j *W.sub.M(s.sbsb.i.sub.)

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j *W.sub.M(s.sbsb.i.sub.) (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, C_S.sbsb.ik is the k-th item in symbol S_i '"'"'s closure and u_s.sbsb.i_n is a position weight.

22. The computer apparatus of claim 17 wherein said means for generating said projection of said first string and said second string uses said cluster tables and said position weight tables by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =d.sub.j *W.sub.M(s.sbsb.i.sub.) *us.sub.i.spsb.n

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j *W.sub.M(s.sbsb.i.sub.) *us.sub.i.spsb.n (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, C_S.sbsb.ik is the k-th item in symbol S_i '"'"'s closure, u_s.sbsb.i_n is a cluster weight, and W_M(s.sbsb.i.sub.) is a position weight.

23. A computer-readable medium having stored thereon a plurality of sequences of instructions for a database application, said plurality of sequences of instructions including sequences of instructions which, when executed by a processor, cause the processor to perform the steps of:

normalizing a first string of S known digitally represented characters to create a first normalized string of N normalized symbols and normalize a second string of known digitally represented characters to create a second normalized string of N normalized symbols;

storing said first normalized string and said second normalized string;

generating a first projection from said first normalized string and a second projection from said second normalized string;

causing said processing means to store said first projection and said second projection;

retrieving said first and second projections from said storage means; and

comparing said first projection and said second projection to determine a degree of similarity of said first and second projections.

24. The computer-readable medium of claim 23 wherein said step of generating said first projection and said second projection uses cluster tables.

25. The computer-readable medium of claim 23 wherein said step of generating said first projection and said second projection uses position weight tables for generating said first projection and said second projection.

26. The computer-readable medium of claim 23 wherein said step of generating said first projection and said second projection uses cluster tables and position weight tables for generating said first projection and said second projection.

27. The computer-readable medium of claim 23 wherein said step of normalizing said first string and said second string generates a medium of a symbol, M, in a normalized image by:

space="preserve" listing-type="equation">M(S.sub.i)=i*|N|/|S|

where S_i is the i-th symbol in a string S, |N| is the normalized size, and |S| is the length of string S.

28. The computer-readable medium of claim 23 wherein said step of generating said projection of said first string and said second string is accomplished by projecting said first string and said second string onto a first and a second closure in a normal distribution by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =d.sub.j

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, and C_S.sbsb.ik is the k-th item in symbol S_i '"'"'s closure.

29. The computer-readable medium of claim 24 wherein said step of generating said projection of said first string and said second string uses said cluster tables by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =d.sub.j *us.sub.i.spsb.n

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j *us.sub.i.spsb.n (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, C_S.sbsb.ik is the k-th item in symbol S_i '"'"'s closure, and u_s.sbsb.i_n is a cluster weight.

30. The computer-readable medium of claim 25 wherein said step of generating said projection of said first string and said second string uses said position weight tables by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =d.sub.j *W.sub.M(s.sbsb.i.sub.)

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j *W.sub.M(s.sbsb.i.sub.) (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, C_S.sbsb.ik is the k-th item in symbol S_i '"'"'s closure and u_s.sbsb.i_n is a position weight.

31. The computer-readable medium of claim 26 wherein said step of generating said projection of said first string and said second string uses said cluster tables and said position weight tables by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =d.sub.j *W.sub.M(s.sbsb.i.sub.) *us.sub.i.spsb.n

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j *W.sub.M(s.sbsb.i.sub.) *us.sub.i.spsb.n (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, C_S.sbsb.ik is the k-th item in symbol S_i '"'"'s closure, u_s.sbsb.i_n is a cluster weight, and W_M(s.sbsb.i.sub.) is a position weight.

32. In a computer system having a processing means coupled to a storage means and a display means, a computer-implemented method of comparing, in a spell checking application, a first string of digitally represented characters with a second string of digitally represented characters, each character of said first and second string of digitally represented characters being a member of a set of digitally represented characters, said computer-implemented method comprising the steps of:

using said processing means, normalizing said first string of S known digitally represented characters to create a first normalized string of N normalized symbols and normalizing said second string of known digitally represented characters to create a second normalized string of N normalized symbols;

storing said first normalized string and said second normalized string in said storage means;

using said processing means, generating a first projection from said first normalized string and a second projection from said second normalized string;

storing said first projection and said second projection in said storage means; and

using said processing means, retrieving said first and second projections from said storage means and comparing said first projection and said second projection to determine a degree of similarity of said first and second projections.

33. The computer implemented method of claim 32 wherein said set of digitally represented characters is a set of ASCII characters.

34. The computer implemented method of claim 32 wherein said set of digitally represented characters is a set of symbols.

35. The computer implemented method of claim 32 wherein said steps of generating said first projection and said second projection further include the step of using cluster tables for generating said first projection and said second projection.

36. The computer implemented method of claim 32 wherein said steps of generating said first projection and said second projection further include the step of using position weight tables for generating said first projection and said second projection.

37. The computer implemented method of claim 32 wherein said steps of generating said first projection and said second projection further include the step of using cluster tables and position weight tables for generating said first projection and said second projection.

38. The computer implemented method of claim 32 wherein said steps of normalizing said first string and said second string further include the step of generating a medium of a symbol, M, in a normalized image by:

space="preserve" listing-type="equation">M(S.sub.i)=i*|N|/|S|

where S_i is the i-th symbol in a string S, |N| is the normalized size, and |S| is the length of string S.

39. The computer-implemented method of claim 32 wherein said step of generating said first projection and said second projection is accomplished by projecting said first string and said second string onto a first and second closure in a normal distribution by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i)+|D|+j =d.sub.j

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, and Cs_ik is the k-th item in symbol S_i '"'"'s closure.

40. The computer implemented method of claim 35 wherein said step of generating said projection of said first string and said second string using cluster tables is accomplished by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =d.sub.j *us.sub.in

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j *us.sub.in (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, C_s.sbsb.ik is the k-th item in symbol S_i '"'"'s closure, and us_in is weight of symbol n in the cluster whose core is S_i.

41. The computer implemented method of claim 36 wherein said step of generating said projection of said first string and said second string using position weight tables is accomplished by

42. The computer-implemented method of claim 37 wherein said step of generating said projection of said first string and said second string using cluster tables and weight tables is accomplished by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =d.sub.j *W.sub.M(s.sbsb.i.sub.) *us.sub.i.spsb.n

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j *W.sub.M(s.sbsb.i.sub.) *us.sub.i.spsb.n (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, C_S.sbsb.ik is the k-th item in symbol S_i '"'"'s closure, u_s.sbsb.i_n is a cluster weight, and W_M(s.sbsb.i.sub.) is a position weight.

43. A computer apparatus for comparing, in a spell checking application, a first string of digitally represented characters with a second string of digitally represented characters, each character of said first and second string of digitally represented characters being a member of a set of digitally represented characters, said computer apparatus comprising:

a storage means;

a processing means coupled to said storage means, said processing means including;

a means for normalizing a first string of S known digitally represented characters to create a first normalized string of N normalized symbols and for normalizing a second string of known digitally represented characters to create a second normalized string of N normalized symbols;

a means for storing said first normalized string and said second normalized string;

a means for generating a first projection from said first normalized string and a second projection from said second normalized string;

a means for storing said first projection and said second projection in said storage means;

a means for retrieving said first and second projections from said storage means; and

a means for comparing said first projection and said second projection to determine a degree of similarity of said first and second projections.

44. The computer apparatus of claim 43 wherein said set of digitally represented characters is a set of ASCII characters.

45. The computer apparatus of claim 43 wherein said set of digitally represented characters is a set of symbols.

46. The computer apparatus of claim 43 wherein said means for generating said first projection and said second projection uses cluster tables.

47. The computer apparatus of claim 43 wherein said means for generating said first projection and said second projection uses position weight tables.

48. The computer apparatus of claim 43 wherein said means for generating said first projection and said second projection uses cluster tables and position weight tables.

49. The computer apparatus of claim 43 wherein said means for normalizing said first string and said second string generates a medium of a symbol, M, in a normalized image by:

space="preserve" listing-type="equation">M(S.sub.i)=i*|N|/|S|

where S_i is the i-th symbol in a string S, |N| is the normalized size, and |S| is the length of string S.

50. The computer apparatus of claim 43 wherein said means for generating said projection of said first string and said second string is accomplished by projecting said first string and said second string onto a first and a second closure in a normal distribution by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =d.sub.j

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, and C_S.sbsb.ik is the k-th item in symbol S_i '"'"'s closure.

51. The computer apparatus of claim 46 wherein said means for generating said projection of said first string and said second string uses said cluster tables by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =d.sub.j *us.sub.i.spsb.n

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j *us.sub.i.spsb.n (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, C_S.sbsb.ik is the k-th item in symbol S_i '"'"'s closure, and u_s.sbsb.i_n is a cluster weight.

52. The computer apparatus of claim 47 wherein said processing means for generating said projection of said first string and said second string uses position weight tables by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =d.sub.j *W.sub.M(s.sbsb.i.sub.)

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j *W.sub.M(s.sbsb.i.sub.) (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, C_S.sbsb.ik is the k-th item in symbol S_i '"'"'s closure and u_s.sbsb.i_n is a position weight.

53. The computer apparatus of claim 48 wherein said means for generating said projection of said first string and said second string uses said cluster tables and said position weight tables by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =d.sub.j *W.sub.M(s.sbsb.i.sub.) *us.sub.i.spsb.n

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j *W.sub.M(s.sbsb.i.sub.) *us.sub.i.spsb.n (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, C_S.sbsb.ik is the k-th item in symbol S_i '"'"'s closure, u_s.sbsb.i_n is a cluster weight, and W_M(s.sbsb.i.sub.) is a position weight.

54. A computer-readable medium having stored thereon a plurality of sequences of instructions for a spell checking application, said plurality of sequences of instructions including sequences of instructions which, when executed by a processor, cause the processor to perform the steps of:

normalizing a first string of S known digitally represented characters to create a first normalized string of N normalized symbols and normalize a second string of known digitally represented characters to create a second normalized string of N normalized symbols;

storing said first normalized string and said second normalized string;

generating a first projection from said first normalized string and a second projection from said second normalized string;

causing said processing means to store said first projection and said second projection;

retrieving said first and second projections from said storage means; and

comparing said first projection and said second projection to determine a degree of similarity of said first and second projections.

55. The computer-readable medium of claim 54 wherein said step of generating said first projection and said second projection uses cluster tables.

56. The computer-readable medium of claim 54 wherein said step of generating said first projection and said second projection uses position weight tables for generating said first projection and said second projection.

57. The computer-readable medium of claim 54 wherein said step of generating said first projection and said second projection uses cluster tables and position weight tables for generating said first projection and said second projection.

58. The computer-readable medium of claim 54 wherein said step of normalizing said first string and said second string generates a medium of a symbol, M, in a normalized image by:

space="preserve" listing-type="equation">M(S.sub.i)=i*|N|/|S|

where S_i is the i-th symbol in a string S, |N| is the normalized size, and |S| is the length of string S.

59. The computer-readable medium of claim 54 wherein said step of generating said projection of said first string and said second string is accomplished by projecting said first string and said second string onto a first and a second closure in a normal distribution by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =d.sub.j

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, and C_S.sbsb.ik is the k-th item in symbol S_i '"'"'s closure.

60. The computer-readable medium of claim 55 wherein said step of generating said projection of said first string and said second string uses said cluster tables by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =d.sub.j *us.sub.i.spsb.n

space="preserve" listing-type="equation">C.sub.S.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j *us.sub.i.spsb.n (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, C_S.sbsb.ik is the k-th item in symbol S_i '"'"'s closure, and u_s.sbsb.n is a cluster weight.

61. The computer-readable medium of claim 56 wherein said step of generating said projection of said first string and said second string uses said position weight tables by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =d.sub.j *W.sub.M(s.sbsb.i.sub.)

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|-j =d.sub.j *W.sub.M(s.sbsb.i.sub.) (j=0,1,2, . . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, C_S.sbsb.ik is the k-th item in symbol S_i '"'"'s closure and u_s.sbsb.n is a position weight.

62. The computer-readable medium of claim 57 wherein said step of generating said projection of said first string and said second string uses said cluster tables and said position weight tables by:

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sbsb.i.sub.)+|D|+j =d.sub.j *W.sub.M(s.sbsb.i.sub.) *us.sub.i.spsb.n

space="preserve" listing-type="equation">C.sub.s.sbsb.i.sub.M(S.sub.i.sub.)+|D|-j =d.sub.j *W.sub.M(s.sbsb.i.sub.) *us.sub.i.spsb.n (j=0,1,2. . . , |D|-1)

where D is a distributing series, |D| is distribution size, d_j is the j-th item in distribution series D, C_S.sbsb.ik is the k-th item in symbol S_i '"'"'s closure, u_s.sbsb.i_n is a cluster weight, and W_M(s.sbsb.i.sub.) is a position weight.