Inductive inference method for obtaining rules represented by propositional logic
First Claim
1. An inductive inference method for obtaining necessary and sufficient conditions represented by propositional logic in a computerized expert system by inducing expert knowledge necessary for knowledge information processing, comprising the steps of:
- (1) reading into the computer an initial necessary condition and an initial sufficient condition stored in a memory of said expert system;
(2) reading example data from said memory, said example data having proposition, conditions of said proposition and tendency of each condition, said proposition and conditions having truth or falsity, said tendency having a first tendency (I), a second (II) or a third (III), said first tendency positively supporting the truth of the proposition, said second tendency positively supporting the falsity of the proposition, and said third tendency not positively supporting the truth or falsity of the proposition;
(3) obtaining a necessary condition and a sufficient condition based on the following result such that when the proposition is false, said necessary condition is obtained by logically ORing said initial necessary condition and said condition having the falsity and said tendency (I), by logically ORing said initial necessary condition and a logically negated condition of said condition having the truth and said tendency (II), by logically ORing said initial necessary condition and said condition having the falsity and said tendency (III), and by logically ORing said initial necessary condition and a logically negated condition of said condition having the truth and said tendency (III), and when the proposition is true, said sufficient condition is obtained by logically ANDing said initial sufficient condition and said condition having the truth and said tendency (I), by logically ANDing said initial sufficient condition and a logically negated condition of said condition and a logically negated condition of said condition having the falsity and said tendency (I), by logically ANDing said initial sufficient condition and said condition having the truth and said tendency (III), and by logically ANDing said initial sufficient condition and a logically negated condition of said condition having the falsity and said tendency (III); and
(4) repeating said steps (2) and (3) for another example data to obtain updated necessary condition and sufficient condition, by using said necessary condition and said sufficient condition as said initial necessary condition and said initial sufficient condition, respectively, in step (3).
1 Assignment
0 Petitions
Accused Products
Abstract
An inductive inference apparatus comprises an input section for inputting a proposition, conditions for the proposition, and the tendency of each condition, a storage section for storing the proposition, and necessary and sufficient conditions of the proposition, a condition detecting section for forming the necessary and sufficient conditions for the truth or falsity of the proposition in accordance with the input proposition, the input conditions, and the input tendency of each condition, a judging section for, with respect to an example in which the truth or falsity of the proposition is unknown, judging the truth or falsity of the proposition using already stored necesary and sufficient conditions, and a control section for, when the truth or falsity of the proposition in a new example input to the input section is known, supplying the conditions of the proposition and the tendency of each condition to the condition detecting section to store the necessary and sufficient conditions formed by the condition detecting section in the storage section, and for, when the truth or falsity of the proposition in the example is unknown, supplying the conditions of the proposition to the judging section.
21 Citations
4 Claims
-
1. An inductive inference method for obtaining necessary and sufficient conditions represented by propositional logic in a computerized expert system by inducing expert knowledge necessary for knowledge information processing, comprising the steps of:
-
(1) reading into the computer an initial necessary condition and an initial sufficient condition stored in a memory of said expert system; (2) reading example data from said memory, said example data having proposition, conditions of said proposition and tendency of each condition, said proposition and conditions having truth or falsity, said tendency having a first tendency (I), a second (II) or a third (III), said first tendency positively supporting the truth of the proposition, said second tendency positively supporting the falsity of the proposition, and said third tendency not positively supporting the truth or falsity of the proposition; (3) obtaining a necessary condition and a sufficient condition based on the following result such that when the proposition is false, said necessary condition is obtained by logically ORing said initial necessary condition and said condition having the falsity and said tendency (I), by logically ORing said initial necessary condition and a logically negated condition of said condition having the truth and said tendency (II), by logically ORing said initial necessary condition and said condition having the falsity and said tendency (III), and by logically ORing said initial necessary condition and a logically negated condition of said condition having the truth and said tendency (III), and when the proposition is true, said sufficient condition is obtained by logically ANDing said initial sufficient condition and said condition having the truth and said tendency (I), by logically ANDing said initial sufficient condition and a logically negated condition of said condition and a logically negated condition of said condition having the falsity and said tendency (I), by logically ANDing said initial sufficient condition and said condition having the truth and said tendency (III), and by logically ANDing said initial sufficient condition and a logically negated condition of said condition having the falsity and said tendency (III); and (4) repeating said steps (2) and (3) for another example data to obtain updated necessary condition and sufficient condition, by using said necessary condition and said sufficient condition as said initial necessary condition and said initial sufficient condition, respectively, in step (3). - View Dependent Claims (2, 3, 4)
-
Specification