JPS61123947A - Intelligence editing method - Google Patents

Abstract

PURPOSE:To realize the easy and natural intelligence editing of various classes of users, by providing interfaces which are designed to match the various images held by users of various levels who utilize the data base system. CONSTITUTION:An end user interface 15 and an expert interface 16 execute interaction which matches the though of an end user 19, using definition 17 or 18 of terminal windows which are designed to match with various images and editors 11-14. Also, an intelligence comprehensive control system 21 checks and administrates imperfect intelligence or contradiction between the intelligences which may occur because of one intelligence being edited by plural number of user interfaces. In this way, users can edit the intelligence complying with their own levels and as the data base system is utilized in the order of its completion, the users of various level can make easy and natural intelligence editing.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a knowledge editing method, and particularly to a knowledge editing method that provides a user interface according to the user's level.

[Conventional technology]

To use a database, specialized knowledge about databases is required. Therefore, in order for users who do not have specialized knowledge about databases to use it, K has decided to interpose a special program in between. For example, the system is configured so that users can easily make inquiries in Japanese. Therefore, the recognition of relationships between data will differ depending on the user's specialized knowledge and the configuration of the system that directly handles the data. In other words, the essence of a database is that a database is logically made up of multiple tables, each table is made up of multiple fields, word knowledge is stored in a word dictionary, and it is also expressed in classes and arcs. Knowledge is expressed in a 7-frame format consisting of slots, 7 assets, Paris, -, etc., and you should understand how to specifically express the knowledge you want to store. are doing. On the other hand, in the case of an end user who queries a database system in Japanese, all knowledge is expressed on the world model, and objects in the target field and relationships between them are expressed as classes and arcs. It is natural to think that words related to these should also correspond to classes or arcbes.

[Problem that the invention seeks to solve]

Conventional knowledge editing systems using databases are configured only for users at a specific level, and there is no system that has the ink 7 ace that is suitable for users at each level, from end users to database experts. There was a problem.

In view of these conventional problems, the present invention aims to provide an interface that matches the different images of knowledge that users at different levels of database use and development have, thereby making it easier for users at all levels to edit knowledge. The aim is to provide a system that can be used naturally.

[Means for solving problems]

According to the present invention, the object is to express knowledge related to the subject field in a world model consisting of a network of classes corresponding to objects and arcs corresponding to relationships between objects, as described in the claims. In a database system that enables database queries in Japanese by assigning words related to each class and arc as well as mapping information for correspondence with the logical structure of the database, users have knowledge of the subject field and knowledge of words. This is achieved by a knowledge editing method characterized in that knowledge regarding the database structure is acquired on the world model by preparing a plurality of ink7aces according to the user's level.

〔Example〕

As an embodiment of the present invention, a knowledge editing system is described that includes two ink7aces: an end-21-the-repel interface and a database expert-level interface. FIG. 1 is a diagram illustrating the level hierarchy of A/7 Ace. At configuration description level 1, objects in the target field and relationships between them are edited using classes and arcs. The content to be edited here is
The relationships between classes by class names and arcs are the same. For example, an editing operation that generates a class called "product" and connects three classes "product name,""product" and "list price" with arcs corresponds to this level. At word assignment level 2, words related to classes or arcs are assigned. At this level, we will deal with content words (nouns, verbs, adjectives) K, including their notation, pronunciation, and parts of speech. For example, the operation of assigning the noun beer to a product corresponds to this level. Composition description level and word assignment level 2
``-cy)'' There is an interface for users to use,
Mapping information corresponding to the class and function words (particles, auxiliary verbs, conjunctions, etc.) unrelated to class arcs are not manipulated. Therefore, even if end users do not know the individual expression formats of each type of knowledge, they can express knowledge in a target field using the two concepts of classes and arcs, and can acquire knowledge by simply assigning words used in that field to classes and arcs. Can be edited. Class content description level 4 handles, in addition to the class information handled at configuration description level 1, the correspondence between the class and the logical structure of the database, and procedures for handling data within the class. At word content description level 3, for all words, including not only content words but also function words, not only the spelling, pronunciation, and part of speech, but also information used for sentence analysis, connection information used for word segmentation, and content words. Information on the corresponding class or arc, and in the case of verbs, case information that it can take, etc. are handled.

The class content description level and the word content description level 3 are interfaces used by database experts (of course, the interface for end users can also be used). The interface is based on the logical structure of dictionaries and databases, and has a format that matches the editing image of experts.

FIG. 2 is a configuration diagram of the knowledge editing system.

Knowledge to be edited includes word dictionary 5, world model 6,
There are three databases 7, and each knowledge has management systems 8 to 10 to manage it. An editor is prepared for each level of user input/removal ace, and each editor 11 to 14 performs editing using the knowledge management system handled by the editor. The end user interface 15 and the expert interface 16 use a terminal window definition 17 or 18 that matches the image of each interface and an editor necessary for that interface to create a dialogue that matches the thinking of the user 19 or 20. I do.

The knowledge integrated management system 21 manages incomplete knowledge, inconsistencies between knowledge, etc. that arise when multiple user interfaces edit one piece of knowledge. For example, there are cases in which the knowledge edited in Engineering/Douser/Re7 Ace 15 can be used as is in a database system, but in other cases it cannot be used until it has been edited at the content description level (for example, due to lack of mapping information). In such cases, the knowledge should be marked as unfinished to prevent the database system from using it unconditionally. The user can later search for this incomplete knowledge and complete the knowledge. Furthermore, at the content description level, since each knowledge of the world model 6, word dictionary 5, and database 7 can be handled individually, the one-time purchase nature of the knowledge may be lost. In the knowledge integrated management system 21, inconsistencies between pieces of knowledge can be checked according to instructions from the user. As a result, users at each level edit knowledge according to their own level, and the database system uses the completed knowledge in order, so when editing knowledge, there is no need to consider inconsistencies in knowledge or completeness for the database system. You can easily collect knowledge RMa.

〔Effect of the invention〕

As explained above, according to the present invention, it is possible to provide inputs suitable for users of different levels of the database system, so that knowledge can be edited easily and naturally, and knowledge can be edited easily and naturally. Changes can also be processed flexibly by users at each level.

Furthermore, since the present invention provides I/7 Ace according to the user's level, it is possible to apply this to check knowledge access rights according to the user's qualifications.

That is, by providing Ink 7 Ace centered on classes and arcs to end users, it is possible to hide knowledge that can be manipulated by experts and prevent unfair manipulation of knowledge.

Furthermore, by providing multiple levels of user interface, it is possible to provide an optimal interface for Bell's users.

[Brief explanation of the drawing]

Figure ta1 is a diagram explaining the hierarchy of interface levels, and Figure 2 is a configuration diagram of the knowledge editing system. 1... Configuration description level, 2... Word assignment level, 3... Word content description level, 4... Class content description level, 5... Word dictionary, 6... World model, 7. ...Database, 8.Dictionary management system, 9.Model management system, 10.Database management system, 11.Configuration description level editor, 2..Word assignment level editor, 13.
... content description level editor, 14... content description level editor, 15... end user interface, 16... extract heart interface, 17... end user level window definition,
18...Expert level window definition,
19...End user, 20...Expert, 21...Knowledge integrated management system (, -1-9゜

Claims (1)

[Claims] The knowledge of the target field is expressed in a world model consisting of a network of classes corresponding to objects and arcs corresponding to relationships between objects, and the correspondence between words related to each class and arc and the logical structure of the database is created. In a database system that allows database inquiries in Japanese by assigning mapping information, multiple interfaces are provided to suit the user's level of knowledge about the target field, vocabulary, and database structure. A knowledge editing method characterized by acquiring on a world model.