Data processing method for a semiotic decision making system used for responding to natural language queries and other purposes

• US 6,278,987 B1
• Filed: 07/30/1999
• Issued: 08/21/2001
• Est. Priority Date: 07/30/1999
• Status: Expired due to Term

• Current Assignee: AutoGnomics Corp., Unisys Corporation
• Original Assignee: AutoGnomics Corp., Unisys Corporation
• Inventors: Zabilansky, Dana M., Waller, John C, Mazzagatti, Jane Campbell, Parker, Charles Austin, McCrory, Duane J., Reed, Frederick W.
• Primary Examiner(s): Chaki, Kakali
• Application Number:
  US09/364,192
• Time in Patent Office: 753 Days
• Field of Search: 712/223, 706/45, 706/46, 706/59, 706/61, 704/1, 704/231, 704/241, 704/251, 704/256, 704/257, 707/3, 707/6, 707/102
• US Class Current: 706/45
• CPC Class Codes: G06F 40/216 using statistical methods
• CPC Class Codes: G06F 40/268 Morphological analysis

Abstract

A semiotic decision making system processes a training corpus of information in the form of sequential sets of elements to create a database which is thereafter used to make decisions relating to queries input in the same type of elements. Sets of sequential elements of a training corpus are received. Ordered pairs of sequential elements and ordered pairs are identified. The ordered pairs include element/element, pair/element, element/pair and pair/pair ordered pairs, in a recursive semiotic process based on the statistical occurrence of element sequences in the training corpus sets whereby each ordered pair represents an n sequential element subset of a training corpus set defined by a set of nested ordered pairs. Constituent sets of ordered pairs and elements are identified for the training corpus sets. Each constituent set for a given training corpus set of n sequential elements has 2n−1 members including representations of each sequential element of the training corpus set and a set of nested subsets of ordered pairs where each sequential training corpus set element is included in one of said ordered pairs. Database records are created corresponding to the identified constituent sets.

Claims

1. A dyadic semiotic processing method for a semiotic decision making system wherein a training corpus of information in the form of sequential sets of elements, where the number of elements of each sequential set does not exceed a selected finite number N, is used to create a database which is thereafter used to make decisions relating to queries input in the same type of elements, the method including the following steps:

receiving sets of sequential elements of a training corpus;

identifying ordered pairs of sequential elements and ordered pairs, said ordered pairs including element/element, pair/element, element/pair and pair/pair ordered pairs, in a recursive semiotic process based on the statistical occurrence of element sequences in the training corpus sets whereby each ordered pair represents an n sequential element subset of a training corpus set defined by a set of nested ordered pairs;

for each training corpus set having at least three elements, identifying constituent sets of ordered pairs and elements, each said constituent set for a given training corpus set of n sequential elements having 2n−

1 members including representations of each sequential element of the training corpus set and a set of nested subsets of ordered pairs where each sequential training corpus set element is included in one of said ordered pairs; and

creating a database records corresponding to said identified constituent sets.

2. A dyadic semiotic processing method according to claim 1 wherein sequential set elements which represent elemental symbols of a training corpus are received.

3. A dyadic semiotic processing method according to claim 2 further comprising the step of:

generating thought signs as said database records from selected constituent sets corresponding to a training corpus set having n elemental symbols by using parenthetical symbols and the sequence n set elements corresponding to said training corpus set where the sequence of set elements is parenthetically grouped in a nested set of n−

1 parenthetical symbol pairs corresponding to the ordered pairs within the constituent set such that within each pair of parenthetical symbols are two ordered items where each item is either a set element or a parenthetical grouping whereby the outermost parenthetical symbol pair contains all of the other parenthetical symbol pairs and all n set elements.

4. A dyadic semiotic processing method according to claim 3 further comprising the step of identifying tokens for processing by a pseudo deduction module by utilizing taxemic dyadic semiotic processing of multiple thought signs corresponding to training corpus sets to identify a most statistically significant thought sign associated with each training corpus set for token identification.

5. A dyadic semiotic processing method according to claim 4 wherein said token identifying step identifies for each training corpus set having at least two elements, the most significant thought sign associated with the training corpus set and that thought sign'"'"'s two ordered items within its outermost parenthetical symbol pair as tokens.

6. A dyadic semiotic processing method according to claim 1 wherein sequential set elements which represent sets of elemental symbols of a training corpus are received.

7. A dyadic semiotic processing method for a semiotic decision making system wherein a training corpus of information in the form of sequential sets of elements, where the number of elements of each set does not exceed a selected finite number N, is used to create a database which is thereafter used to make decisions relating to queries input in the same type of elements, the method including the following steps:

receiving sets of sequential elements of a training corpus;

identifying ordered pairs of sequential elements and ordered pairs, said ordered pairs including element/element, pair/element, element/pair and pair/pair ordered pairs, in a recursive semiotic process based on the statistical occurrence of element sequences in the training corpus sets whereby each ordered pair represents a sequential element subset of a training corpus set defined by a set of nested ordered pairs;

storing data representations of said ordered pairs in a knowledge base such that said knowledge base data representations comprising predicates and elemental and non-elemental acts wherein;

each predicate is associated with a class of one or more acts such that each act is associated with only one class of acts;

each elemental act represents a training corpus set element and defines a single act class of a corresponding elemental predicate; and

each non-elemental act represents a sequence of a case predicate followed by a result predicate, such that all non-elemental acts are recursively defined as representations of at least one set of sequential elements which is a subset of a training corpus set and each predicate represents the sets of sequential elements represented by each act within its associated class of acts;

for each training corpus set having at least three elements, identifying constituent sets of acts representing training corpus sets, each said constituent set representing a given multi-element training corpus set of n sequential elements consisting of 2n−

1 acts including the n elemental acts which represent each sequential element of the given training corpus set and n−

1 non-elemental acts having case and result predicates where each predicate'"'"'s class includes an act within the constituent set thereby defining a nested set of acts having a highest order act representing the entire n sequence of elements; and

creating a database record corresponding to said identified constituent sets.

8. A dyadic semiotic processing method according to claim 7 wherein sequential set elements which represent elemental symbols of a training corpus are received.

9. A dyadic semiotic processing method according to claim 8 further comprising the step of:

generating thought sign denotations as said database records from constituent act sets corresponding to a training corpus set having n elemental symbols using the constituent acts and parenthetical symbols, where the parenthetical symbols represent case and result predicates of non-elemental acts such that the denotation is represented by a nested parenthetical structure where each non-elemental act is followed by a pair of parenthetical symbols within which are the act'"'"'s corresponding case predicate class'"'"' act followed by the act'"'"'s corresponding result predicate class'"'"' act whereby the highest order act is followed by a parenthetical symbol pair which contains all of the other acts and parenthetical symbol pairs.

10. A dyadic semiotic processing method according to claim 9 further comprising the steps of:

receiving sets of performance data after thought sign denotations have been generated for training corpus sets;

identifying a most statistically significant constituent set of acts corresponding to each set of performance data such that said constituent set of acts corresponds to a largest sequential subset of elements of the performance set data which is capable of being represented by the acts stored in said knowledge base;

generating a performance data set denotation from a corresponding most statistically significant constituent set of acts using the constituent acts and parenthetical symbols, where parenthetical symbols represent case and result predicates of non-elemental acts, such that the denotation is represented by a nested parenthetical structure where each non-elemental act is followed by a pair of parenthetical symbols within which are the act'"'"'s corresponding case predicate class'"'"' act followed by the act'"'"'s corresponding result predicate class'"'"' act whereby the highest order act is followed by a parenthetical symbol pair which contains all of the other acts and parenthetical symbol pairs; and

identifying a thought sign denotation corresponding to each set of performance data by comparing the denotation generated for the performance data set with the thought sign denotations generated for the training corpus sets through a scoring process where the score for each training corpus set thought sign denotation is given by;

11. A dyadic semiotic processing method according to claim 10 wherein the thought sign denotation identifying step identifies the thought sign denotation having the highest score as the thought sign denotation identified with the performance data set if the highest score exceeds a given threshold and, if not, identifies the performance set denotation as the thought sign denotation.