Adaptive knowledge base

• US 6,064,971 A
• Filed: 06/07/1995
• Issued: 05/16/2000
• Est. Priority Date: 10/30/1992
• Status: Expired due to Fees

• Current Assignee: William J. Hartnett
• Original Assignee: William J. Hartnett
• Inventors: Hartnett, William J.
• Primary Examiner(s): HAFIZ, TARIQ R
• Application Number:
  US08/476,483
• Time in Patent Office: 1,805 Days
• Field of Search: 707/4, 707/10, 705/10, 705/7, 395/200.49, 348/12, 706/45, 706/46
• US Class Current: 706/46
• CPC Class Codes: G06Q 40/02 Banking, e.g. interest calculation or account maintenance credit or loans G06Q40/03
• CPC Class Codes: G06Q 99/00 Subject matter not provided for in other groups of this subclass

Abstract

A method to operate a computerized adaptive knowledge base. The contents and organization of the adaptive knowledge base evolve based on the contributions and evaluations of a user community. Whether a particular set of contents or an alternative organization is preserved for future iterations is a function of user evaluations, taking into account the amounts of information and the relative importance of content vis-a-vis organization. Items of the adaptive knowledge base which are as yet unevaluated by a particular user are ranked by estimating that user'"'"'s evaluations, based on other items for which evaluations are available to compare with other users. For example, evaluations are estimated from regressions with other users, weighted by functions of the correlation coefficients, the number of observations in common, and the significance of those observations for the particular user in the sense that their evaluation differs from the mean evaluation.

Claims

1. A method of operating a computerized adaptive knowledge base which evolves over a plurality of iterations, evolution of the adaptive knowledge base being based on user evaluations and on user proposed contributions, comprising the steps of:

executing software on the computer to receive from multiple users proposed contributions to the adaptive knowledge base;

executing software on the computer to collect from multiple users evaluations of organization and contents of the adaptive knowledge base, including the proposed contributions;

storing in the adaptive knowledge base the evaluations collected;

associating the evaluations collected with the organization and the contents evaluated;

computing total evaluation functions based upon the evaluations collected;

storing in the adaptive knowledge base the total evaluation functions computed for one of the plurality of iterations;

associating the total evaluation functions computed with the organization and the contents evaluated; and

selecting those portions of the adaptive knowledge base to preserve in a next iteration, based on the total evaluation functions.

2. The method of claim 1, wherein the evaluations include a quality score equal to the average of user evaluations and a quantity score equal to the amount of information and an evaluation function computed includes a predetermined extensivity parameter and is defined as:

(quality score).sup.(1-extensivity) ×

(quantity score)^extensivity

and wherein the evaluation functions computed include content evaluations and organizational evaluations, and a total evaluation function computed includes a predetermined content parameter and is defined as;

{(content)×

(content evaluation)}+{(1-content)×

(organization evaluation)}.

3. A method of managing a data base using a computer which provides a user access to entries therein based on estimates of the user'"'"'s own evaluations, comprising the steps of:

controlling the computer with software to correlate in a memory in the computer the user'"'"'s past actual evaluations of entries with evaluations of those entries by other users; and

controlling the computer with software to estimate the user'"'"'s evaluations for an entry for which the user has not entered an actual evaluation on the basis of the correlations in the memory.

4. A method of managing a data base using a computer which provides a user access to entries therein based on estimates of the user'"'"'s own evaluations, comprising the steps of:

controlling the computer with software to perform one of maximum likelihood estimation analysis, Tobit analysis, logit analysis, probit analysis and regression analysis of the user'"'"'s past actual evaluations of entries against the evaluations of those entries by other users; and

controlling the computer with software to estimate the user'"'"'s evaluations for an entry for which the user has not entered an actual evaluation on the basis of the analysis performed.

5. The method of claim 4, wherein the step of controlling to estimate further comprises the steps of:

storing in a memory an evaluation array E_ij, where i is a user index and j is an observation index;

storing in a memory a weight array W_uj =1+|E_uj -E_j |.sup.α

, where u is a value of the user index representative of the user, α

is a user-specified parameter having a default value of 0, and E_j is an average of evaluations for observation j; and

for each observation j lacking an evaluation by the user, operating the computer to compute;

##EQU3## r_j is a correlation coefficient for a user i from a linear regression of E_uj against E_ij with weights W_ij, where W_ij =W_uj if there is an observation for E_ij and W_ij =0 otherwise, β

is a user-specifiable parameter having a default value of 1, γ

is a user-specifiable parameter having a default value of 1, (m_i E_ij +b_i) is an estimate of E_uj calculated from coefficients m_i and b_i of the linear regression of E_uj against E_ij with weights W_ij, and the summations Σ

_i≠

u include all users i for which evaluation E_ij is present and for which a number of observations in common with user u to perform the linear regression of E_uj against E_ij with weights W_ij exceeds a predetermined threshold.

6. A method of managing a data base using computers, wherein the data base is distributed by segments to a plurality of computers operated by individual users based on their expressions of interest in different components of the data base, comprising the steps of:

controlling each of the plurality of computers to permit the user to tag portions of the data base with indicators of interest;

controlling each of the plurality of computers to permit the user to specify the quantity of information from the data base to be received in response to a query; and

distributing segments of the data base to each of the plurality of computers in accordance with a probabilistic function of the user'"'"'s indicators of interest and quantities specified.

7. A computerized adaptive knowledge base system, wherein the adaptive knowledge base evolves over a plurality of iterations, evolution of the adaptive knowledge base being based on user evaluations and on user proposed contributions, comprising:

a computer executing software for receiving from multiple users proposed contributions to the adaptive knowledge base;

the computer further executing software for collecting from multiple users evaluations of organization and contents of the adaptive knowledge base, including the proposed contributions;

means for storing in the adaptive knowledge base the evaluations collected;

means for associating the evaluations collected with the organization and the contents evaluated;

means for computing evaluation functions based upon the evaluations collected;

means for storing in the adaptive knowledge base the evaluation functions computed for one of the plurality of iterations;

means for associating the evaluation functions computed with the organization and the contents evaluated; and

means for selecting those portions of the adaptive knowledge base to preserve in a next iteration, based on the evaluation functions.

8. The system of claim 7, wherein the evaluations include a quality score equal to the average of user evaluations and a quantity score equal to the amount of information and the evaluation function computed includes a predetermined extensivity parameter and is defined as:

(quality score).sup.(1-extensivity) ×

(quantity score)^extensivity

and wherein the evaluations include content evaluations and organizational evaluations, and the evaluation function computed includes a predetermined content parameter and is defined as;

{(content)×

(content evaluation)}+{(1-content)×

([origination] organization evaluation)}.

9. A system for managing a data base using a computer which provides a user access to entries therein based on estimates of the user'"'"'s own evaluations, comprising:

means for correlating in a memory in the computer the the user'"'"'s past actual evaluations of entries with evaluations of those entries by other users;

means for estimating the user'"'"'s evaluations for an entry for which the user has not entered an actual evaluation on the basis of the correlations in the memory.

10. A system for managing a data base using a computer which provides a user access to entries therein based on estimates of the user'"'"'s own evaluations, comprising:

means for performing one of maximum likelihood estimation analysis, Tobit analysis, logit analysis, probit analysis and regression analysis of the user'"'"'s past actual evaluations of entries against the evaluations of those entries by other users; and

means for estimating the user'"'"'s evaluations for an entry for which the user has not entered an actual evaluation on the basis of the analysis performed.

11. A system for managing a data base using computers, wherein the data base is distributed by segments to a plurality of computers operated by individual users based on their expressions of interest in different components of the data base, comprising:

means for controlling each of the plurality of computers to permit the user to tag portions of the data base with indicators of interest;

means for controlling each of the plurality of computers to permit the user to specify the quantity of information from the data base to be received in response to a query; and

means for distributing segments of the data base to each of the plurality of computers in accordance with a probabilistic function of the user'"'"'s indicators of interest and quantities specified.

12. A method of operating a computerized adaptive knowledge base which evolves over a plurality of iterations, evolution of the adaptive knowledge base being based on user evaluations and on user proposed contributions, comprising the steps of:

executing software on the computer to receive from multiple users proposed contributions to the adaptive knowledge base;

executing software on the computer to collect from multiple users evaluations of portions of the adaptive knowledge base, including the proposed contributions;

storing in the adaptive knowledge base the evaluations collected;

associating the evaluations collected with the portions evaluated; and

selecting those portions of the adaptive knowledge base to preserve in a next iteration, based on the evaluations.

13. A method of operating a computerized adaptive knowledge base which evolves over a plurality of iterations, evolution of the adaptive knowledge base being based on user evaluations and on user proposed contributions, comprising the steps of:

executing software on the computer to receive from multiple users proposed contributions to the adaptive knowledge base, including both contents and organization;

executing software on the computer to collect from multiple users evaluations of the contents and alternative organizations of the adaptive knowledge base, including the proposed contributions;

storing in the adaptive knowledge base the evaluations collected;

associating the evaluations collected with the contents and organizations evaluated; and

selecting those contents and alternative organizations of the adaptive knowledge base to preserve in a next iteration, based on the evaluations.

14. A method of operating a computerized data base with a plurality of users, by which users can adapt the keyword access, comprising the steps of:

providing users the capability to select as a starting keyword template any already available set of keywords and any associated structure;

providing users the capability to modify said keyword template by specifying additional keywords, where the default initialization of whether an entry in the data base matches the additional keywords is able to be specified as a function of the frequency of patterns of words or phrases in said entry, with said default initialization able to be manually overridden for specific entries in the data base;

providing users the capability to modify said keyword template by changing any associated structure; and

providing users the capability to make the adapted keyword access available to other users.