Classification apparatus

• US 6,266,656 B1
• Filed: 09/21/1998
• Issued: 07/24/2001
• Est. Priority Date: 09/19/1997
• Status: Expired due to Fees

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Ohno, Kazuhiko
• Primary Examiner(s): Powell, Mark
• Assistant Examiner(s): Khatri, Anil
• Application Number:
  US09/157,315
• Time in Patent Office: 1,037 Days
• Field of Search: 706/20, 706/47, 706/61, 706/48, 706/62, 706/58
• US Class Current: 706/20
• CPC Class Codes: G06F 16/353 into predefined classes
• CPC Class Codes: G06F 18/24765 Rule-based classification
• CPC Class Codes: G06N 5/025 Extracting rules from data

Abstract

A classification apparatus for performing effective learning type automatic classification for realistic problems of classification. The apparatus includes input unit for entering the known case data and the unknown case data, a classification ruled database for storing classification rules including the probabilistic information, a case database for storing the known case data in the form of a network based on the logical relation of conditional parts, a probability value estimating unit for estimating probability values of the results of classification using the conditional parts of the known case data and the unknown case data as entered and the rules of classification and a classification rule generating unit for evaluating the validity of the classification rules by statistic verification for suppressing generation of useless classification rules, and a negative condition searching unit for receiving all or part of the conditional parts of the known case data as entered. The classification rule generating unit has an added function of generating a classification rule including the negative condition using the negative condition searching unit.

Claims

1. A classification apparatus in which classification rules are automatically generated using known case data, having known results of classification, among case data each of which is made up of a set of a conditional part serving as a clue for classification and a result of classification, and in which unknown case data, having unknown results of classification, are automatically classified using said classification rules;

comprising;

(a) an input unit for entering the known case data and the unknown case data;

(b) a classification rule database for storing classification rules including probabilistic information;

(c) a case database for storing the known case data in the form of a network based on the logical relation of the conditional parts;

(d) a probability value estimating unit for estimating probability values of the results of classification using the conditional parts of the known case data and the unknown case data as entered and the rules of classification;

(e) a classification rule generating unit for evaluating the validity of the classification rules received from the classification rule database and the case database by statistic verification for suppressing generation of useless classification rules, and generating a valid classification rule, wherein the generated valid classification rule comprises at least a value indicating a frequency of the validity of the conditional part; and

(f) an output unit for outputting the generated classification rule, wherein the generated classification rule is used to classify the unknown case data.

2. The classification rule generating apparatus as defined in claim 1 wherein the probability value output from said probability value estimating unit and is the result of classification of said known case data entered from said input unit, and is compared with the probability value of the result of classification output by said probability value estimating unit for generating the classification rule if the difference exceeds a pre-determined value.

3. The classification rule generating apparatus as defined in claim 1 further comprising:

negative condition searching means for receiving all or part of the conditional parts of the known case data as entered into said input unit, said negative condition searching means searching in said case database for inferring case nodes which include the negative conditions;

said classification rule generating means having an added function of generating a classification rule including the negative condition using the negative condition searching means.

4. A recording medium having recorded thereon a program to be executed by a computer, said program comprising:

(a) processing of generating a new classification rule on the entered known case data from the classification rule database in the classification apparatus as defined in claim 1, and

(b) processing of evaluating rule validity by statistic verification for suppressing generation of wasteful rules.

5. A recording medium having recorded thereon a program to be executed by a computer as defined in claim 4, wherein said program comprises:

(a) processing of searching said case database using all or part of the entered known case data in the recording medium for inferring case nodes including the negative condition by way of negative condition retrieving processing, and

(b) processing of generating a classification rule including a negative condition inferred by said searching processing of said negative condition by way of generating the classification rule having the negative condition.

6. A computer system in which classification rules are automatically generated using known case data, having known results of classification, among case data each of which is made up of a set of a conditional part serving as a clue for classification and a result of classification, and in which unknown case data, having unknown results of classification, are automatically classified using said classification rules;

comprising;

an input unit for entering the known case data and the unknown case data;

a classification rule database for storing classification rules including probabilistic information; and

a case database for storing the known case data in the form of a network based on the logical relation of the conditional parts;

a recording medium having recorded thereon a program for allowing the following means to function on said computer system;

(a) means for estimating probability values of the results of classification using the conditional parts of the known case data and the unknown case data as entered and the rules of classification; and

(b) means for evaluating the validity of the classification rules by statistic verification to suppress generation of useless classification rules;

(c) means for generating classification rules using said known case data and said unknown case data, wherein the generated classification rules comprises at least a value indicating a frequency of the validity of the conditional part;

(d) an output unit for outputting the classification rules,

wherein the generated classification rule is used to classify the unknown case data.

7. A computer system in which classification rules are automatically generated using known case data, having known results of classification, among case data each of which is made up of a set of a conditional part serving as a clue for classification and a result of classification, and in which unknown case data, having unknown results of classification, are automatically classified using said classification rules;

comprising;

an input unit for entering the known case data and the unknown case data;

a classification rule database for storing classification rules including probabilistic information; and

a case database for storing the known case data in the form of a network based on the logical relation of the conditional parts;

a recording medium having recorded thereon a program for allowing the following means to function on said computer system;

(a) means for estimating probability values of the results of classification using the conditional parts of the known case data and the unknown case data as entered and the rules of classification;

(b) negative condition searching means for receiving, as input, all or part of the conditional parts of the known case data as entered, and

(c) classification rule generating means for evaluating the validity of the classification rules by statistic verification to suppress generation of useless classification rules, and generating a valid classification rule, wherein the generated valid classification rule comprises at least a value indicating a frequency of the validity of the conditional part; and

said negative condition searching means searching in said case database for inferring case nodes which include the negative conditions, said classification rule generating means having the function of generating the classification rule including the negative condition using said negative condition searching means; and

(f) an output unit for outputting the classification rules,

wherein the generated classification rule is used to classify the unknown case data.