Fuzzy retrieval apparatus and method

• US 5,583,966 A
• Filed: 03/14/1994
• Issued: 12/10/1996
• Est. Priority Date: 10/04/1991
• Status: Expired due to Term

• Current Assignee: U Krowel LLC (Intellectual Ventures LLC)
• Original Assignee: Omron Corporation
• Inventors: Nakajima, Hiroshi
• Primary Examiner(s): Downs, Robert W.
• Assistant Examiner(s): Hafiz, Tariq R.
• Application Number:
  US08/204,336
• Time in Patent Office: 1,002 Days
• Field of Search: 395/3, 395/10, 395/11, 395/50, 395/51, 395/61, 395/600, 395/900, 395/147, 395/934
• US Class Current: 706/52
• CPC Class Codes: G06F 16/2468 Fuzzy queries
• CPC Class Codes: Y10S 706/90 Fuzzy logic
• CPC Class Codes: Y10S 706/934 Information retrieval or Information management

Abstract

The present invention provides an apparatus and method in which it is possible to set information retrieval conditions wherein the user'"'"'s intention in performing retrieval is reflected satisfactorily, and in which fuzzy retrieval can be performed in accordance with the retrieval conditions thus set. According to the present invention, the user, in the setting of retrieval conditions, is capable of freely connecting preconditions, which represent the desired conditions with regard to attribute signals in a data base, using one or a plurality of MEAN, MIN and MAX operations. Accordingly, it is possible to set retrieval conditions that fully reflect the intentions of the user. Moreover, fuzzy retrieval processing in accordance with retrieval conditions set by the user is executed and retrieved results reflecting the intentions of the user are obtained.

Claims

1. A fuzzy retrieval apparatus comprising:

a data base in which a plurality of types of attribute signals with respect to an item of a plurality of items are registered in advance for each item;

means for setting signals representing a retrieval condition which comprises at least two preconditions each indicating a desired condition relating to at least one of the plurality of types of the attribute signals, and a connecting condition that connects these preconditions;

membership operating means for producing degree-of-membership signals that indicate the degrees to which an attribute signal conforms to set preconditions;

MEAN operating means for producing and outputting a signal indicating a mean value of a plurality of applied degree-of-membership signals;

MIN operating means for selecting and outputting a signal, which indicates a minimum value, from a plurality of applied degree-of-membership signals;

MAX operating means for selecting and outputting a signal, which indicates a maximum value, from a plurality of applied degree-of-membership signals, and

control means for applying the degree-of-membership signals obtained from the membership operating means to at least one of the MEAN operating means, MIN operating means and MAX operating means in accordance with a predetermined connecting condition, and outputting a degree-of-concurrence signal obtained from a final stage of at least one of the MEAN operating means, MIN operating means and MAX operating means.

2. A fuzzy retrieval apparatus according to claim 1, wherein said means for setting retrieval conditions comprises input means for entering retrieval-condition signals.

3. A fuzzy retrieval apparatus according to claim 1, wherein said means for setting retrieval conditions comprises means for storing a plurality of retrieval-conditions signals in advance, and means for selecting at least one of said retrieval-condition signals from the retrieval-condition signals that have been stored.

4. A fuzzy retrieval apparatus according to claim 1, wherein said membership operating means produces said degree-of-membership signal for each item with regard to all preconditions that have been set, and said control means controls said at least one of the MEAN operating means, MIN operating means and MAX operating means in such a manner that a degree-of-concurrence signal is obtained for each and every item.

5. A fuzzy retrieval apparatus according to claim 1, wherein said control means applies an output signal from said MEAN operating means, said MIN operating means or said MAX operating means to at least one of MEAN operating means, MIN operating means and MAX operating means in accordance with the predetermined connecting condition.

6. A fuzzy retrieval apparatus according to claim 1, further comprising means for producing a plurality of weight signals, each weight signal representing a weight of a degree-of-membership signal;

said MEAN operating means, MIN operating means and MAX operating means performing operations corresponding to a MEAN operation, MIN operation and MAX operation, respectively, with regard to said plurality of weight signals applied together with the degree-of-membership signals; and

said control means applying said plurality of weight signals to at least one of MEAN operating means, MIN operating means and MAX operating means together with the degree-of-membership signals, and outputting a signal, which represents degree of importance obtained from the final stage of said at least one of MEAN operating means, MIN operating means and MAX operating means, together with the degree-of-concurrence signal.

7. A fuzzy retrieval apparatus according to claim 6, wherein said means for producing a plurality of weight signals comprises a data base in which, together with the attribute signals, there are registered signals representing degrees of credibility of the attribute signals, means for entering a signal, which represents degree of attached significance, for each attribute, and multiplying means for multiplying a signal representing degree of credibility and a signal representing degree of attached significance together and outputting a weight signal.

8. A fuzzy retrieval apparatus according to claim 6, wherein said means for producing a plurality of weight signals is a data base in which, together with the attribute signals, there are registered weight signals representing degrees of credibility of the attribute signals.

9. A fuzzy retrieval apparatus according to claim 6, wherein said means for producing a plurality of weight signals is means for entering, for each attribute, a weight signal representing degree of attached significance of the attribute.

10. A fuzzy retrieval method, comprising the steps of:

registering, in a data base, in advance for each item of a plurality of items, a plurality of types of attribute signals with respect to an item of said plurality of items;

setting signals representing a retrieval condition comprising a plurality of preconditions indicating desired conditions relating to at least one of the plurality of types of attribute signals, and a connecting condition that connects these preconditions;

producing signals representing degrees of membership that indicate the degrees to which an attribute signal conforms to a predetermined precondition;

executing at least one of a MEAN operation, MIN operation and MAX operation in accordance with a predetermined connecting condition and with regard to the degree-of-membership signals obtained from membership operating means, wherein the MEAN operation is for producing a signal indicating a mean value of the degree-of-membership signals, the MIN operation is for selecting a signal, which indicates a minimum value, from the degree of membership signals, and the MAX operation is for selecting a signal, which indicates a maximum value, from the degree-of-membership signals, and

outputting a signal representing the degree of concurrence obtained by executing the operation.