Patent
1993-12-22
1997-10-21
Black, Thomas G.
395605, 395 61, G06F 1730
Patent
active
056806060
DESCRIPTION:
BRIEF SUMMARY
TECHNICAL FIELD
This invention relates to a fuzzy retrieval apparatus and method, as well as an apparatus for creating membership functions.
BACKGROUND ART
Fuzzy retrieval is a retrieval method that allows fuzziness of data stored in a data base or fuzziness of retrieval conditions.
Though a conventional fuzzy retrieval apparatus allows the setting or entry of retrieval conditions by fuzzy information (e.g., a fuzzy language), data stored in a data base is limited to definite numerical values (crisp numbers) or information.
Fuzzy retrieval in the conventional fuzzy retrieval apparatus is executed in the following manner: A given retrieval condition is expressed by a membership function. From the data that has been stored in a data base, the data that corresponds to the retrieval condition is read out and the degree of membership of the data that has been read out in association with the membership function representing the retrieval condition is calculated. Processing for calculating degree of membership is executed with regard to all retrieval conditions and with regard to all data corresponding to the retrieval conditions. Data having fairly good conformity to the retrieval conditions is eventually selected from the viewpoint of whether the value of a degree of membership thus obtained is highest, greater than a predetermined value or relatively high.
In a case where it is attempted to register data (fuzzy data) having fuzziness in a data base, the fuzzy data must be converted into crisp data in the conventional fuzzy retrieval apparatus. For example, with regard to fuzzy information "young in age", an operation is performed in which this is registered in a data base upon converting it into the crisp data "15 to 25 years of age".
According to conventional fuzzy retrieval, data conforming to given retrieval conditions is fetched from the data base as mentioned above. However, the user must make a judgment regarding the results of retrieval thus obtained. For example, it is necessary for the user himself to observe the data of the retrieved results and then make a judgment as to why such retrieved results were obtained and what action should be taken with respect to these retrieved results. If such decision making is delayed, there may be instances in which the reason for performing the retrieval is rendered meaningless. For example, if such a delay occurs in the search of a customer data base, speedy action of the user with regard to a customer is delayed and a promising customer may fail to be acquired.
In fuzzy reasoning processing which includes the above-described fuzzy retrieval processing, use of a membership function in order to express the fuzzy information is unavoidable. Creation of the membership function generally is performed on the basis of experimentation or experience. There are also situations in which a new membership function is created based upon a membership function that already exists. In either case, the creation of the membership function is carried out by hand. The problem that arises is expenditure of much labor and time. In addition, when a membership function is created by hand, the membership function created reflects a difference among individuals. This leads to a problem in that it is not always possible to attain a standard basis for making judgments.
DISCLOSURE OF THE INVENTION
An object of the present invention is to make it possible to store fuzzy information in a data base in its own form without converting the fuzzy information into crisp data.
Another object of the present invention is to make it possible to save both definite crisp information and vague fuzzy information in a data base in mixed form.
A further object of the present invention is to provide a system capable of assisting a user in decision making with regard to results of fuzzy retrieval.
A further object of the present invention is to provide an apparatus that utilizes an existing membership function to create a new membership function automatically and in a short period of time, thereby reduci
REFERENCES:
patent: 5267348 (1993-11-01), Someya et al.
patent: 5422984 (1995-06-01), Iokibe et al.
patent: 5467270 (1995-11-01), Matsumoto
patent: 5471559 (1995-11-01), Yasunobu et al.
patent: 5471677 (1995-11-01), Imanaka
APS Text Search & Retrieval Classroom Manual, USPTO, May 1987, pp. 2-37 through 2-39.
Lee, J.H. et al. "Enhancing the Fuzzy Set Model for High Quality Document Rankings." Microprocessing & Microprogramming, vol. 35, No. 1/5, Sep. 1992, pp. 337-344.
Nomura, T. et al. "Generating Ambiguous Attributes for Fuzzy Queries." IEEE International Conference on Fuzzy Systems, 12 Mar. 1992, San Diego, CA, pp. 753-760.
Bosc, P. et al. "Flexible Selection Among Objects: A Framework Based on Fuzzy Sets." ACM SIGIR 11th International Conference on Research & Development in Information Retrieval, 15 Jun. 1988, pp. 433-449.
Zemankova, M. "Filip: A Fuzzy Intelligent Information System with Learning Capabilities." Information Systems, vol. 14, No. 6, 1989, pp. 473-486.
Kacprzyk, J. et al. "FQUERY 111+: A `Human Consistent` Database Querying System Based on Fuzzy Logic with Linguistic Quantifiers." Information Systems, vol. 14, No. 6, 1989, pp. 443-453.
"Fuzzy Database Retrieval and Manipulation Language," Electronic Communication Institute Technical Report, Jan. 27, 1976, Shingakugiho, vol. 78, No. 233, pp. 65-72.
"Perspectives and Trends of Fuzzy Databases," Electronic Communication Institute Technical Report, Jan. 27, 1976, Shingakugiho, vol. 78, No. 233, pp. 207-214.
"A Relational Database Extended By Application Of Fuzzy Set Theory And Linguistic Variables," Computers and Artificial Intelligence, vol. 8, No. 2, 1989, Le. T. Young, et al., pp. 153-168.
"Perspectives of Fuzzy Databases," Motohide Umano, et al., Department of Precision Engineering, Faculty of Engineering, Osaka University, vol. 3, No. 1, 1991, pp. 2-14.
"A Fuzzy Database System For Ambiguous Queries," Takahiko Nomura, et al., Faculty of Science and Technology, Keio University, 1991, pp. 91-98.
Hayashi Motoji
Kudo Toshimi
Nakajima Hiroshi
Black Thomas G.
Omron Corporation
Von Buhr Maria N.
LandOfFree
Apparatus and method for creating membership functions does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Apparatus and method for creating membership functions, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Apparatus and method for creating membership functions will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-1015436