Fuzzy reasoning database question answering system

• US 5,384,894 A
• Filed: 06/11/1993
• Issued: 01/24/1995
• Est. Priority Date: 05/16/1991
• Status: Expired due to Term

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Kobrosly, Walid, Triantafyllos, George, Vassiliadis, Stamatis
• Primary Examiner(s): Downs, Robert W.
• Application Number:
  US08/076,372
• Time in Patent Office: 592 Days
• Field of Search: 395/61, 395/900, 395/934, 395/600
• US Class Current: 706/52
• CPC Class Codes: G06F 16/2468 Fuzzy queries
• CPC Class Codes: G06F 16/3334 Selection or weighting of terms from queries, including natural language queries
• CPC Class Codes: G06F 16/334 Query execution G06F16/335 takes precedence
• CPC Class Codes: G06F 18/232 Non-hierarchical techniques
• CPC Class Codes: G06F 8/20 Software design
• 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 706/934 Information retrieval or Information management
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access

Abstract

A system provides question answering based on fuzzy logic. The proposed system provides the capability to assess if a database contains information pertinent to a subject of interest by evaluating each comment in the database via a fuzzy evaluator which attributes a fuzzy membership value indicating its relationship to such subject. An assessment is provided for the database as a whole regarding its pertinence to the subject of interest, and consequently comments that are considered as irrelevant to the subject may be discarded. The system has been developed for the examination of databases and evaluated by testing against those which were created for assessing whether computer system development databases contain information pertinent to the functional changes that would occur in the development cycle. The system can be applied with minimal changes to a variety of circumstances provided that the fundamental assumptions for the development of the membership functions are respected for the application.

Claims

1. A question answering system for evaluating a database having text record entries, comprising:

a computer having processors and a storage system which processes instructions and a data in original database storage means for storing an original database, said computer system having

unique word processing means for generating unique words,

unique word storage means for storing a list of unique words,

a replace table generator for generating a replace table,

replace table storage means for storing the replace table,

a word processor,

modified database storage means for storing a modified database, and delete table storage means for storing a delete table,

a unique word generator for generating relevant words,

relevant word table storage means for storing a relevant word table,

a fuzzy expected value (FEV) processor,

combined storage means for storing a relevant word table with a confidence vector,

a fuzzy evaluator and a confidence vector analyzer,

query means for a user entry of a requested subject of interest,

said query means initiating system operation to retrieve data records for evaluation by said system,

said fuzzy evaluator then automatically evaluating the retrieved data records to determine whether it pertains to the requested subject of interest and passing a judgment to the system user in the form of a confidence value that represents the possibility that the retrieved data records do indeed deal with the requested subject of interest,

wherein the data records have text comments written in a spoken language and

wherein said fuzzy evaluator includes a result analyzer for finding the comments of said database that have confidence values greater than a value set by a user of the system.

2. A question answering system for evaluating a database having text record entries, comprising:

a computer having processors and a storage system which processes instructions and a data in original database storage means for storing an original database, said computer system having

unique word processing means for generating unique words,

unique word storage means for storing a list of unique words,

a replace table generator for generating a replace table,

replace table storage means for storing the replace table,

a word processor,

modified database storage means for storing a modified database, and delete table storage means for storing a delete table,

a unique word generator for generating relevant words,

relevant word table storage means for storing a relevant word table,

a fuzzy expected value (FEV) processor,

combined storage means for storing a relevant word table with a confidence vector,

a fuzzy evaluator and a confidence vector analyzer,

query means for a user entry of a requested subject of interest,

said query means initiating system operation to retrieve data records for evaluation by said system,

said fuzzy evaluator then automatically evaluating the retrieved data records to determine whether it pertains to the requested subject of interest and passing a judgment to the system user in the form of a confidence value that represents the possibility that the retrieved data records do indeed deal with the requested subject of interest,

wherein said fuzzy expected value (FEV) processor includes means for summarizing the results of interviews with experts as one input and an input from said relevant word table storage means and said processor is controlled for assigning a fuzzy expected value to all words in a relevant word table, and

wherein the system includes means for assigning a membership value to each word in the relevant word table by taking the average of all interview responses for each word.

3. A question answering system for evaluating a database having text record entries, comprising:

a computer having processors and a storage system which processes instructions and a data in original database storage means for storing an original database, said computer system having

unique word processing means for generating unique words,

unique word storage means for storing a list of unique words,

a replace table generator for generating a replace table,

replace table storage means for storing the replace table,

a word processor,

modified database storage means for storing a modified database, and delete table storage means for storing a delete table,

a unique word generator for generating relevant words,

relevant word table storage means for storing a relevant word table,

a fuzzy expected value (FEV) processor,

combined storage means for storing a relevant word table with a confidence vector,

a fuzzy evaluator and a confidence vector analyzer,

query means for a user entry of a requested subject of interest,

said query means initiating system operation to retrieve data records for evaluation by said system,

said fuzzy evaluator then automatically evaluating the retrieved data records to determine whether it pertains to the requested subject of interest and passing a judgment to the system user in the form of a confidence value that represents the possibility that the retrieved data records do indeed deal with the requested subject of interest,

wherein said fuzzy expected value (FEV) processor includes means for summarizing the results of interviews with experts as one input and an input from said relevant word table storage means and said processor is controlled for assigning a fuzzy expected value to all words in a relevant word table,

wherein the one input resulting from interviews with experts provides confidence values produced by the interviews and this input and the records of the modified database are inputs to said fuzzy expected value processor, and

wherein the system includes means for assigning a confidence value to a data record, and if a record of the modified database contains two words, then the system assigns the result of the function;

##EQU17## to the confidence of that record, and wherein in said function w_i and w_j are the confidence of two words, and wherein "k" is a number greater than zero (0<

k<

7) and "1" is a number between 0 and 1 (1=0.5).

4. A computer system for processing a database having language text records to evaluate their relevance to a predetermined subject of interest, the system comprising:

processors for processing instructions and the text records;

database storage means for storing an original text database;

a unique word generator for generating and storing a list of unique words contained in a text database;

a relevant word generator for generating and storing a list of relevant words including a system user'"'"'s selections from a list of unique words of the original text database generated and stored by the unique word generator;

a modified database generator for generating and storing a modified database, the modified database including the list of relevant words and synonyms, if any, associated with each word in the list of relevant words, whereinafter the unique word generator generates a list of unique words of the modified database;

a relevant word table including the list of unique words of the modified database;

means for calculating and storing confidence values associated with each word in the relevant word table, the confidence values based on input values from personnel other than the system user that reflects their perceptions of a relevance of each word in the relevant word table to the subject of interest; and

a fuzzy expected value (FEV) processor for determining and storing a fuzzy expected value for each word in the relevant word table, the fuzzy expected value based on the confidence values associated with each word in the relevant word table.

5. The system according to claim 4, further comprising:

a fuzzy evaluator for calculating and storing a joined confidence value of each record in the original text database, the joined confidence value of each record based on the fuzzy expected value for each word in that record.

6. The system according to claim 5, wherein the fuzzy evaluator calculates the joined confidence value of a record containing zero words to be equal to zero.

7. The system according to claim 5, wherein the fuzzy evaluator calculates the joined confidence value of a record containing one word to be equal to the fuzzy expected value of the one word.

8. The system according to claim 5, wherein the fuzzy evaluator calculates the joined confidence value of a record containing two words according to the formula ##EQU18## where w_i and w_j are fuzzy expected values of the two words;

k is a constant greater than zero;

1 is a value greater than zero and less than one; and

c is equal to (1-m)/m where m is an average value of the two words.

9. The system according to 5, wherein the fuzzy evaluator calculates the joined confidence value of a record containing three or more words according to the formula

10. The system according to claim 5, wherein the system further comprises means for retrieving a record having a joined confidence value greater than a predetermined value.

11. A computer system for processing a database having language text records to evaluate their relevance to a predetermined subject of interest, the system comprising:

processors for processing instructions and the text records;

database storage means for storing an original text adatbase;

a unique word generator for generating and storing a list of unique words contained in a text database;

a relevant word generator for generating and storing a list of relevant words including a system user'"'"'s selections from a list of unique words of the original text database generated and stored by the unique word generator;

a modified database generator for generating and storing a modified database, the modified database including the list of relevant words and synonyms, if any, associated with each word in the list of relevant words, whereinafter the unique word generator generates a list of unique words of the modified database;

a relevant word table including the list of unique words of the modified database;

means for calculating and storing confidence values associated with each word in the relevant word table, the confidence values based on input values from personnel other than the system user that reflects their perceptions of a relevance of each word in the relevant word table to the subject of interest; and

wherein the unique word generator generates and stores a list of all unique words contained in the text database.

12. A computer system for processing a database having language text records to evaluate their relevance to a predetermined subject of interest, the system comprising:

processors for processing instructions and the text records;

database storage means for storing an original text database;

a unique word generator for generating and storing a list of unique words contained in a text database;

a relevant word generator for generating and storing a list of relevant words including a system user'"'"'s selections from a list of unique words of the original text database generated and stored by the unique word generator;

a modified database generator for generating and storing a modified database, the modified database including the list of relevant words and synonyms, if any, associated with each word in the list of relevant words, whereinafter the unique word generator generates a list of unique words of the modified database;

a relevant word table including the list of unique words of the modified database;

means for calculating and storing confidence values associated with each word in the relevant word table, the confidence values based on input values from personnel other than the system user that reflects their perceptions of a relevance of each word in the relevant word table to the subject of interest; and

wherein the unique word generator includes means for calculating and storing a frequency of each unique word contained in the text database.

13. A method for evaluating a relevance of a database having text records to a predetermined subject of interest, the method comprising computer executed steps of:

compiling and storing a list of unique words contained in the database having text records;

extracting and storing from the list of unique words a list of relevant words selected by a user of the method;

storing confidence values furnished by personnel other than the user who are familiar with the subject of interest that reflect their perceptions of a relevance of each word in the list of relevant words to the subject of interest; and

calculating and storing a fuzzy expected value for each word in the list of relevant words, the fuzzy expected value for each word based on the confidence values for each word in the list of relevant words.

14. The method according to claim 13, further comprising a computer executed step of calculating a joined confidence value of each text record of the database based on fuzzy logic that incorporates the fuzzy expected value of each word in the text record.

15. The method according to claim 14, further comprising a computer executed step of retrieving a text record of the database having a joined confidence value within a predetermined range.

16. The method according to claim 14, wherein the step of calculating a joined confidence value of each text record includes the steps of:

determining a paired confidence value for all pairs of adjacent words in a text record containing more than two words according to the formula

space="preserve" listing-type="equation">n-1

space="preserve" listing-type="equation">Φ

μ

.sub.A [(w.sub.i w.sub.i+1)]

space="preserve" listing-type="equation">i=1

and

assigning a highest paired confidence value of all the pairs of adjacent words in the record as the joined confidence value of the record if the highest paired confidence value exceeds a predetermined upper threshold.

17. The method according to claim 16, wherein the step of assigning a highest paired confidence value includes the step of assigning a lowest paired confidence value of all the pairs of adjacent words in the record as the joined confidence value of the record if none of the paired confidence values exceed the upper threshold and the lowest paired confidence value is lower than a predetermined lower threshold.

18. The method according to claim 17, wherein the step of assigning a lowest paired confidence value includes the step of assigning the average of the fuzzy expected values of all the words in the text record as the joined confidence value of the text record if none of the paired confidence values is lower than the lower threshold nor higher than the upper threshold.

19. The method according to claim 14, wherein the step of calculating a joined confidence value of each text record is based on the formula ##EQU19## for records containing two words, where w_i and w_j are fuzzy expected values of the two words;

k is a constant greater than zero;

1 is a value greater than zero and less than one; and

c is equal to (1-m)/m where m is an average fuzzy expected value of the two words.

20. A computer system for a user to process a database having text data records to evaluate their relevance to a predetermined subject of interest, the system comprising:

means for selecting a list of relevant words from the text data records;

means for personnel to store and assign confidence values to each word in the list of relevant words that represent their opinion of a relevance of the word to the subject of interest;

means for retrieving text data records containing words assigned confidence values satisfying a predetermined formula; and

means for determining and storing a fuzzy expected value for each word in the list of relevant words, the fuzzy expected value based on the confidence values assigned to each word in the list of relevant words.