JPH02141828A - Method for supporting definition of correspondence between expert system and data base system and knowledge processing system using the same - Google Patents

Abstract

PURPOSE:To allow the tables of a relational data base and the classes of an expert system to correspond to each other by expressing the schema of a table selected from the relational data base with an entity-relation set. CONSTITUTION:A user interface 101 is provided for interface between a user 102 and a knowledge processing system and performs the operation of the expert system and a correspondence definition supporting system 103. The step where the user 102 selects tables required in the expert system out of the tables of the relational data base and the step where each table is regarded as one entity set to determine the attribute having the frame name of the expert system in accordance with the attribute of the entity set are performed for all tables selected by the user 102 to generate a model using entity sets and relation sets, and each set is expressed as a class frame.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention provides a method for supporting correspondence definition between an expert system and a relational database system, and a knowledge processing system using the same, that is, a frame and relational The correspondence with data stored in a database system relates to a method for supporting definition and a knowledge processing system using the same.

[Conventional technology]

In this type of conventional technology, the logical structure of the database and the knowledge representation of the expert system are directly correlated. For example, in mapping tables and frames in a relational database as shown in Figure 2, the schema of the table is mapped to the class frame, and one tuple (tuple: a combination of data in one row of the table) in the table is mapped to the expert system. is associated with one instance frame. In addition, as a specific example of representing a tuple of a relational database table in an instance frame in this way, the Information Processing Society of Japan
5th (late 1986) National Conference, Collection of Lectures (■) p.
``Expert system construction tool ES/X90 (8) Knee database knowledge conversion function'' described in p. 1747-1748 is known.

[Problem to be solved by the invention] In the above-mentioned conventional technology, it is difficult to determine what kind of conceptual model the data in the database is constructed with, that is, what kind of "things" and "relationships between things" they represent. Because the data in existing databases is treated as a frame with its logical structure, users write rules keeping in mind "objects" and "relationships between objects," and create a knowledge processing system. The problem was that it needed to be built.

The present invention has been made in view of the above circumstances, and its purpose is to solve the above-mentioned problems in the conventional technology, to clearly grasp the "objects" and "relationships between objects" that the database has, and to A correspondence definition support method between an expert system and a relational database system for determining the correspondence between relational database data and expert system frames, and a knowledge processing system using the same, that is, knowledge processing using the above support method. In order to do this, a correspondence definition support system in which a user maps data in a relational database with frames of an expert system on a computer, a function to link the correspondence definition support system with an expert system, and an inference to determine commands related to the database. The object of the present invention is to provide a knowledge processing system having a function of converting data from a database and a frame of an expert system.

[Means to solve the problem]

The above object of the present invention is to construct a specific structure on secondary storage,
In a method that supports the definition of the correspondence between data in a relational database to be managed and frames of an expert system that constructs a knowledge base on main memory and performs inference using an inference engine, a user can a step of selecting an item required by the system, a step of determining an attribute having a frame name of the expert system from the attributes of the entity set by considering each table as one entity set, and an attribute having no frame name. a step of determining to which attribute information the information possessed belongs;
Then, the step of dividing the entity set using this information to create multiple entity sets and relationship sets connecting them is performed for all the tables selected by the user, and by integrating sets with the same name. A method for supporting definition of correspondence between an expert system and a database system, characterized by creating a model using entity sets and relationship sets, and representing each set as a class frame; and a method for supporting definition of correspondence between an expert system and a database system as described above. is a knowledge processing system using an interface with a user, a conversion definition support system that associates database data with frames of an expert system based on the entity set and relationship set, and frames and rules. an expert system having a stored knowledge base and an inference engine that references frames within the knowledge base and executes rules; a database management system that manages the database; A combination interface between the database and the expert system that performs the interface processing with the expert system, and a system that converts frames in the file system into a format that can be stored in the knowledge base, sends it to the database-expert system combination interface, and stores it in the knowledge base. This is achieved by a knowledge processing system characterized by having a frame translator and a database for converting from a form to a text format and storing it in a file system, and a file system for storing files.

[Effect]

In the method for supporting correspondence definition between an expert system and a database system according to the present invention, the schema of a table selected from a relational database is represented by an entity-relationship set, and an attribute having a representative name of the entity-relationship is determined.

By representing the entity-relationship set as a class frame of the expert system, it is possible to indirectly associate the table of the relational database with the class of the expert system, and the field having the name of the instance frame belonging to the class is determined.

Note that there is an entity-relationship model as a method of replacing database data with a conceptual model.

As shown in Figure 3 later, this entity-relationship model is based on a rectangular set of entities (301, 303) and a set of relations (3
02) can be represented as a diamond and the set can be graphed by connecting the set with a line. Regarding this method, A
CM Transactions on Databa
seSyste+is, vol, 1. &1. pp, 1
9 (March, 1976) is helpful.

It has also been proposed that the attributes of each set can be visually understood using a graph in which a description representing the attributes is added to the entity-relationship set. Regarding this, see Sakai: "Information resource management techniques - database design using the ER model"
(Ohmsha, 1987.pp, 31) is helpful.

〔Example〕

Embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a diagram showing the configuration of a knowledge processing system that is an embodiment of the present invention. In FIG. 1, a user interface 101 is an interface between a user 102 and a knowledge processing system, and is a system for operating an expert system and a correspondence definition support system, which will be described later. The correspondence definition support system 103 is a system that supports correspondence between a table schema and a class frame in order to use data in a database as a frame for an expert system. The inference engine 104 is a system that interprets and executes the knowledge held in the knowledge base 105. Further, the knowledge base 105 is an area that secures knowledge in the main memory, and has rules (referring to "knowledge representing rules") and frames (referring to "knowledge representing facts") as knowledge.

The database-expert system coupling interface 106 is a system for using data in a database as a frame in an expert system.

Database management system (DBMS) 107
is a system that manages data in a database. The frame translator 108 is a system that mutually converts frames in the file system 111 and frames in the knowledge base 105. Translator 1
09 is a system that converts frames and rules in the file system 111 into a format that can be stored in the knowledge base 105 and stores it. The database 110 is a system having a secondary storage area that stores data in a fixed format. The file system 111 is a system that stores data on secondary storage.

In the knowledge processing system shown in FIG.
constructs knowledge using the user interface 101 and stores it in the file system 111. Furthermore, the correspondence definition support system 103 is used to determine the correspondence between data in the database and knowledge, and the correspondence definition is stored in the file system 111. Through the user interface 101, the user 102 sends the data →
Converting data in the file system 111 into a frame and storing the frame in the database system 110 using the corresponding definition in the file system 110 by sending a frame conversion, frame → data conversion command; data in the database 110; can be converted into a frame and stored in the file system 111.

When the user 102 accesses the knowledge in the file system 111,
When data→frame conversion, frame→data conversion instructions, and database access instructions are written, knowledge and correspondence definitions are stored in the knowledge base 105 using the translator 109, and the inference engine 104 is started, the knowledge base 105
The inference engine 104 interprets and executes the contents. Note that at this time, the inference engine 104 converts the data frame in the knowledge.

Frame→data conversion commands and database access commands are determined, the commands are sent to the database-expert system coupling interface 106, the data in the database 110 is converted into frames using the corresponding definition in the knowledge base 105, and the data is converted into the knowledge base. Converting frames in the knowledge base 105 into data and storing them in the database system 110
10 searches and updates can be performed.

FIG. 2 is a diagram showing an example of the data structure of a database used in the knowledge processing system. This database is a relational database, where each table is represented in a tabular format and each table has a name. A table has fields, and each field has a name. In FIG. 2, the company table 201 includes r company name J, r number of employees,
The person table 202 has a field called “Capital”.
is "Mr."

It has fields ``1 person'', ``special skill'', and ``company name''. Also, tables have tuples. A tuple is a set of data in one row. This database also has tuple search and update commands.

FIG. 3 shows an entity-relationship set used in the display of the correspondence definition support system 103 of the knowledge processing system. The entity sets 301 and 302 are displayed as rectangles, and the relationship set 303 is displayed as a diamond. Each set has 30 attribute names
Has 4. Entity sets 301 and 302 are sets of entities having the same characteristics, and a relation set 303 is a set of relationships between the entity sets. Connecting the entity set and the relationship set with a line represents the connection between the relationship set and the entity set.

FIG. 4 is an execution flow of the knowledge processing system. When the knowledge processing system is started, an initial menu display 401 is displayed on the display, and a process selection 402 is performed. When the start of the correspondence definition support system is selected, the expert system performs a necessary table selection process 403, and then performs a table-frame correspondence definition process 404.When the process 404 is completed, a text editor execution process 403 is performed.
05, the corresponding definition is displayed on the display. When the correspondence definition is completed, the process returns to process selection 402. When you select another process in the process selection 402, the process selection 40 is selected again.
Do step 6. When starting a text editor is selected, knowledge is created and modified through text editor execution processing 405. When starting an expert system is selected, inference processing 407 is performed. After execution of inference process 407, process selection 4
Return to 02.

FIG. 5 is a flowchart of the process of selecting tables required by the expert system. When the table selection process begins, a list 501 of tables available to the user is displayed, and then process selection 502 is performed. When the table selection process is selected, the system waits for user input. From the displayed table view, the user can use the mouse with a button to
When one necessary table is selected (process 503), the table selected by the user is made to blink (process 504). After the table blinks, the process returns to process selection 502. Processing selection 502
If you select table search, the system will wait for user input, and when the user inputs a command (process 505),
The database is searched and the results are displayed on the display (process 506). After displaying the results, process selection 502
Return to

FIG. 6 is a flowchart of the table-frame correspondence definition process. When the correspondence definition process begins, an entity set having the table name and field name selected in the previous table process is displayed on the display (process 601). The state waits for user input for table selection, and the user selects an undivided entity set using a mouse with a button (process 602).
. Next, the state waits for attribute selection input, and when the user selects an attribute with a frame name used in the expert system using a mouse with a button (process 603), the attribute name, entity ID, and relationship ID on the entity set model are enters the input state to set the , and sets each person by user input (
Process 604).

After that, a determination process is performed in which the user is asked to input whether or not to continue the selection process of the attribute with the frame name (process 60
5) If the selection process of an attribute having a frame name is continued, the state is set to wait for user input for attribute selection. If the selection process for attributes having a frame name is not continued, the process moves to the selection process for attributes not having a primary frame name. When the process moves to selecting an attribute that does not have a frame name, the state waits for user input for attribute selection, and when the user selects an attribute using a mouse with a button (process 606), the attribute name on the entity set model and the entity ID , the input state is entered to set any of the related IDs, and each is set by user input (processing 607).

At this time, the user must not set the entity ID and the relationship ID to one attribute at the same time. This is because if attributes are set to one attribute at the same time, it becomes unclear to which set the attribute will be assigned when dividing the entity set. after that,
A determination process 608 is performed in which the user is asked to input whether or not to continue the attribute selection process, and if the attribute selection process is continued, the state waits for user input for attribute selection, and if the selection process is not continued, the entity set division process and A set name is set on the model (process 609). After executing process 609,
A determination process 610 is performed in which the user is asked to input whether process 609 has been performed for all entity sets, and if process 609 has not been performed for all entity sets, process 602 is performed. Processing for all entity sets 6
If step 09 has been performed, sets with the same name are integrated on the model (process 611).

FIG. 7 is a flowchart of processing 609 in FIG.

When process 609 starts, in process 701 new entity sets are created as many as the entity IDs of the entity set to be divided into one, and in process 702 new relationship sets are created as many as the number of relationship IDs. Next, extract attributes with the same entity ID from the entity set to be divided and set them in the new entity set.
3) Give the set entity set the same name as the original entity set (process 704).

Next, it is determined whether or not there is an unset entity set in the new entity set (process 705), and if there is an unset entity set, the process returns to process 703, and if there is no unset entity set, the process moves to the next process. .

Next, create a key item by extracting attributes with frame names with the same relationship ID from the entity set to be divided, extract attributes that do not have frame names with the same relationship from the entity set to be divided, and create a new Process 706) to set the relationship set, connect a line with the newly created entity set that has the same relationship ID, and give the set relationship set the same name as the original entity set (process 707). next,
It is determined whether or not there is an unset relation set in the new relation set (process 708). If there is an unset relation set, the process returns to step 706, and if there is no unset relation set, the process moves to the next process.

Next, the entity set of the split lights is deleted (processing 709).

At this point, the entity set after division is displayed on the display. Furthermore, the user input state for entity set selection is entered, and the newly created set is selected using a mouse with a button (process 710). When a set is selected, the set name on the model is entered, and the user sets the set name (
Processing 711). Next, a determination is made to have the user input whether or not to set a group name (process 712), and if the group name is to be set, the process returns to process 710, and if a group name is not to be set, process 609 ends. do.

FIG. 8 is a flowchart of processing 611 in FIG. 6.

In process 611, new entity sets are created by the number of types of entity set names on the model (process 801) L, and new relationship sets are created by the number of types of relationship set threads on the model (process 802). Next, as the entity set name on the model, attributes of sets with the same name are grouped together and set as a new unset entity set (process 803), and the same name as the original entity set is given (process 4804).

If there is an unset entity set among the newly created entity sets, the process returns to process 803, and if there is no unset entity set, the process moves to the next process (process 805).

Next, combine the attributes of the sets that have the same name as the relational prefecture name on the model, and set it to a new relational set that has not been set (
Process 806) L, give it the same name as the original relation set (process 807), and connect the newly created entity set with the same name as the entity set related to the original relation set with a line (process 808).
. If there is an unset relation set among the newly created relation sets, the process returns to process 806, and if there is no unset relation set, the process moves to the next process (process 809). next,
The original set and line are deleted (process 810). At this point, the integrated collection appears on the display.

The internal table 901 in FIG. 9 is referenced and updated by the processing such as setting the entity set base in FIGS. 6, 7, and 8. One row of the internal table 901 represents information on attributes that the set has. Rows in internal tables are added when a set is created and deleted when the set is deleted.

Each column of the internal table has the following information.

``Set I DJ'' 902 is set at the time of entity set creation in processes 601, 701, etc. The value of ``collective yarn'' 903 is determined by setting the collective yarn on the model in process 711. The “table name” 904 includes the process 60
1, the name of the table selected in the table selection process is set. “Attribute name” 905 includes processing 604.
In process 607, the value input by the user is set 6
The field value of the table selected in the table selection process in process 601 is set in the "field name" 906.

“Entity I DJ 907 and “Relationship I DJ 90g
D is processing 604. In process 607, the value input by the user is set. "Settings" 909 does not have a value at the time the set is created, and the "Settings" 909 does not have a value at the time the set is created. When the attribute is set on the set by operation 706, the value It 171 is set.

"New" 910 is the process 701. The value It 177 is set when creating a new set in process 702, etc., and process 709. All values are deleted when deleting the original set in process 810.
Change to

The “type of set” 911 is determined by the process 701. When creating a new entity set in process 702, the value (tOjJ) is set for the entity set, and the value 14117 is set for the relationship set.The value LtO''' is set for "key determination" 912 when creating a new entity set, and the value is set in process 603. If you select an attribute that has a frame name, the value will be changed to #J 171. “Associated entity IDJ9
13 sets the set ID of the entity set connecting the lines in process 707.

The “key item” 914 is set by extracting the “attribute name J 905. The contents of the internal table 901 represent the states of 1208.1209.1210 in FIG. 12, which will be shown later.

10, 11, and 1 show examples of the display when the processes in FIGS. 5, 6, 7, and 8 are executed.
This will be explained using Figure 2. Screen 1001 in FIG.
is a table list display screen. 1005 is a process selection command. "Select", "Explore".

A process can be selected by moving the mouse pointer 1008 to the character string "modeling" and pressing the mouse button. Also, when selecting a table, move the mouse pointer 10 to the table part displayed on the screen.
You can select by moving 08 and pressing the mouse button. Screen 1002 is a database search screen. Reference numeral 1006 is a process selection command, and similarly to the above, a process can be selected using the mouse pointer 1008. "
Command>> When box 009 is selected, an input field is created in 1009, allowing the user to enter a search command.

Further, one screen 1003 is a modeling support screen for correspondence definition processing. Reference numeral 1007 is a process selection command, and similarly to the above, a process can be selected using the mouse pointer 1008.

FIGS. 11 and 12 are model display examples of correspondence definition support processing. The table used in this example is the table shown in FIG. 2 above. 1101 and 1102 are initial displays, which are displayed on the display by processing 601. 1103 is a display when selecting an entity set. Points to the entity set to be selected using the mouse pointer 1008. 1104 is a display when selecting an attribute having a frame name and setting entity ID and relationship ID. An attribute having a frame name is selected by pointing the attribute name with a mouse pointer 1008 as shown in 1104.

This allows the attribute name input field 11 on the model to be
06. Entity ID input field 1107. Relationship ID input fields 1108 appear, and the user sets values in each input field. Symbol 1105 indicates an attribute that has a frame name when each person is set. In 1104, the attribute "Mr.", 1 person" is set as an attribute that has a frame name.
, r company name" is set. 1109 is a display when selecting other attributes and setting entity ID and relationship ID. Mouse pointer 1 on the attribute name, such as 1109
The attribute is selected by pointing at 008. This allows attribute names and entity IO on the model.

Since input fields for relationship IDs appear, the user sets each person in each input field. 1109 indicates that the attribute "special skills" has the name "special skills of r people" on the model and is an attribute of an entity having the entity ID (1).

1201, 1202, and 1203 are displays when the entity set 1109 is divided using relation sets. 12
01 represents an entity set with entity ID (1), and 120
3 represents an entity set with entity ID (2), 1202
represents a relationship set with relationship ID (1). 1204.
1205 and 1206 are displays when setting a set name on the model. As shown in 1206, a set is selected by pointing at the set name with the mouse pointer. This creates a field 120 for inputting the set name on the model.
Since 7 appears, the user sets a value in input field 1207. 1208.1209.1210 is a display after integration of sets with the same name on the model. 12
08 is an entity set with the name “r company”, ! 209
is a relation set with the name “employment”, and 1210 is “
It represents an entity set with the name "Employee".

FIG. 13 is a description of a frame representing a correspondence definition.

The correspondence definition is created from the internal table shown in FIG. Of these descriptions, the portions excluding the underlined portions are the same as the description of the correspondence definition sent to the text editor of the user interface after the correspondence definition process is completed. Note that the underlined portion cannot be determined by the correspondence definition process, so after being sent to the text editor, the user edits it using the text editor. The description in FIG. 13 will be explained below.

In the figure, "Conversion definition IJ1301 is a frame used when converting data in the database into a frame".
There are 3 named slots, each slot representing the following information: rclassJ 1302 indicates that "conversion definition 1" is an instance frame of a class called "conversion definition frame." Next, ``User ID DJ1303'' represents the user identification name when using the database, ``Password J1304 represents the password when using the database. 1305 is the name of the table containing the data to be converted to a frame.
In Conversion Definition I J1301, "r company table".

The data of "One Person Table" is used as a frame.

This uses the value of "table name" 904 in FIG.

"Key slot name J1306 is the name of a slot with a frame name, and "key field name J1307 is a field name of a table with a frame name.

“Key slot name J1306 and “key field name” 1
The order of the values 307 has is related to 6, that is.

In the case of "conversion definition I J1301," the value of the field "r company name" in the "company table" is used as the value of the slot "r company name." this is.

This is the relationship between the "attribute name" 905 and the "field name" 906 of the attribute whose "table name" 904 in FIG.

[Slot name J1308 is the name of a slot other than the slot with a frame name, and is the name of a slot other than the slot with a frame name.
09 is the name of a field other than the field having a frame name. Between “slot name” 130B and “field name J 1309”, “key slot name” 130
6 and "Key field name J1307. In other words, in [Conversion definition I J1301], the value of "Number of employees" of the "Company table" is used as the value of the slot "Number of employees". This has "value of table name J1305 in r table name" 904 in FIG.
This is the relationship between the “r attribute name of the attribute having the value 1401#” 905 in the “key determination” 912 and the “field name” 906.

“Upper class name” 1310 is the name of the class to which the frame into which the data extracted from the table was converted belongs.

The frame converted using the conversion definition IJ1301 belongs to the "r company" class. This is the "set group" in Figure 9.
This is the relationship between 903 and “table name” 904.

The r frame type J1311 is "upper class name J131
0 indicates whether the frame of the class specified is an entity or a relationship. This is the relationship between "set group" 903 and "set type" 911 in FIG. "Conversion definition 1" 1
In step 301, the frame of the class ``R Company'' designated as the upper class is an entity.

“Search conditions” 1312 are conditions for retrieving data from the table. ``The order of values held by search condition J1312 is related to ``table name J1304.''In conversion definition IJ1301, neither r company table'' nor ``person table'' has a search condition, ``file name J1313
"File name J1313 represents the file name in which the frame is stored when data in the database is converted to a frame through the user interface and stored in the file system." has no meaning, ``In conversion definition IJ1301, there is no file that you want to store.''

On the other hand, [Storage Definition I J1314 is a frame used when converting a frame to database data. Each slot represents the following information.

"Storage Definition I rclassJ130 that J1314 has
2. rUserIDJ1303. r password J130
4 has the same meaning as "conversion definition IJ1301."Storage definition IJ1314 is an instance frame of the storage definition frame."Table name J131" in FIG.
5 is the name of the table that stores the frame, and is the value of "table name" 904 in FIG. [Storage definition I J1
At 314, the frame is stored in the "r company" table. “Key slot name J1316 has the name of the slot that has the frame name of the frame to be stored.

[Key field name J1317] has the name of the field that stores the frame name of the frame to be stored. [“Key slot name” 130 of conversion definition IJ1301
6 and ``key field name'' 1307, there is a relationship between the values of ``key slot name J 1316 and [key field name J 1317''. This means that the value of ``table name'' 1315 is This is the relationship between the "attribute name" 905 and the "r field name" 906 of the attribute which has the "table name" 904 and has the value at 1 t+ in the "key determination" 912.

"Slot name J1318 has the name of the slot that the frame to store has.""Field name J1319 has.

It has the name of the field that stores the slot value.

Also between "slot name J1318 and "field name J1319.

"Key slot name J1316" and "Key field name J1
There is a similar relationship with 317. This is [table name J13
15 in the "table name" 904 in Figure 9, and "
Key determination" 912 is the relationship between the "attribute name" 905 and the "field name" 906 of the attribute that has the value ((OIt). The "upper class name J 1320 has the name of the class to which the frame stored in the table belongs. This is the relationship between "set group" 903 and "table name" 904 in FIG.

The r frame type J1321 indicates whether the type of frame stored in the table is an entity or a relationship. There is a relationship similar to name J1307.This is the relationship between "set group" 903 and "set type" 911 in FIG. "File name J1322" is a file name in which a frame is stored when a frame in a file system is converted to data in a database through a user interface.
2J has no meaning when a frame→data conversion instruction is executed from the inference engine. "Storage definition I J
At 1314, there is no file with the frame that you want to store in the database.

These conversion definitions are created by the number of sets created in the correspondence definition process, and storage definitions are created by the number of tables selected in the correspondence definition process. FIG. 14 is a description of the class frame generated from the internal table created by the correspondence definition. Of these descriptions, the portions excluding the underlined portions are similar to the description of the class frame that is sent to the text editor of the user interface after the correspondence definition processing is completed. Since the underlined portion cannot be determined by the correspondence definition process, it is sent to the text editor and then edited by the user using the text editor. FIG. 14 will be explained below.

The r company J1401 is a class frame and has a name that exists in the "collection name" 903 in FIG. rguper
-class J1402 is "r company" is rsystem
“$Company name J”, which indicates that it is a subclass of J.
1403 is a slot with a frame name. “$
"Company Name" 1403 is "Key Judgment" 912 in FIG.
has the value Kl 1 jj, which is the value of the "attribute name" 905. This slot has string in its data type, "Number of Employees J1404.r Company Capital 41405 is
It's a slot. ``Number of employees J1404゜r Company capital J
1405 is the value of the "attribute name" 905 which has the value "0" in the "key determination" 912 in FIG. "Number of employees" 1404
, r company's capital J1405 both have int as the data type, "Employee J 1406 is r company J1
Similar to 401, this is a class frame that has the value of "attribute name" 905 in FIG. 9 as a slot.

r$Last name J1407. r$ name J1408 C is an attribute with the value "1" in "Key judgment" 912 in FIG.
Human special skill J 1409 has value II in "key judgment" 912
It has an attribute of O17. [Employment J1410 is a class frame.

In FIG. 9, "Attribute name J 905. r key judgment" 912 of the row with the value "Employment" in "Collection name" 903
However, since the set "Employment" is known to be a set that expresses a relationship by the "Set Type" 911, the "Key Judgment" of the set with the ID of "Related Entity I DJ 913" is the value Kl I Hl to the slot with the frame name of [Working J1410].

As a result, "Employment 41410" is changed to "$Company name 41
411゜r$surname” 1412. "$Name" has 1413 slots.

Note that a slot with a "$" mark at the beginning of the slot name is a slot with a frame name and cannot be rewritten.

Figure 15 shows the conversion definition I J130 shown in Figure 13.
1 and ``Conversion Definition Frame'' and ``Storage Definition Frame'' which are classes of ``Storage Definition I J1314''. "
Conversion definition frame J1501 contains “conversion definition IJ130
It has all the slots explained in 1.However, since this frame is a class frame, the slot rclass
instead of J, rsuper-classJ1502
rSuper-class J1502 with “
This indicates that "conversion definition frame" is a subclass of "rsysten+". rqueueJ1503 is a characteristic that slots have, and a slot with the description rqueueJ1503 can have multiple values, and the value that is given to the slot first is taken out first when it is taken out. , [Storage definition frame J1504 has all the slots explained in "Storage definition I J1314. However, since this frame is also a class frame, instead of slot rclassJ, rsupe
It has r-class J1502.

16 and 17, a method for using database data in an expert system will be explained.

FIG. 16 explains how to execute an instruction to convert database data into a frame and a frame to database data from the user interface 101 of the knowledge processing system.

In the knowledge processing system of this embodiment, when the above command is executed from the user interface 101, the database 11
Convert the data in 0 to a frame and save it to file system 1
11. It also converts frames stored in the file system 111 into data and stores them in the database 111.
Store in.

1601 is an expert system screen, and the commands of the expert system are laid out on the screen 1601. The user 102 selects the command by pointing the mouse pointer 1008 at this command.

To convert data in the database 110 into a frame and store it in the file system 111, use command 1.
603 with the mouse pointer.

Then, a prompt 1604 is displayed on the screen 1601. User 102 enters a file name in input area 1605 following prompt 1604 . This file name is the name of the file in which the conversion definition is stored. Similarly, when converting a frame stored in the file system 111 into data and storing it in the database 110, point the command 1602 with the mouse pointer. At this time as well, a prompt 1604 is displayed on the screen, so the user 102 inputs the name of the file in which the storage definition is stored.

Figure 17 shows the inference engine 10 that the knowledge processing system has.
From No. 4 onwards, there are instructions for converting database data into frames, instructions for converting frames into database data, and methods for executing instructions for accessing the database. In order to execute the above instruction from the inference engine 104, the knowledge base 105 needs to have rules and correspondence definitions having the above instruction.

In order to have rules and correspondence definitions in the knowledge base 105, the file system 111 has a file that describes the rules and correspondence definitions, and the translator 10
9 causes the knowledge base 105 to read the rules and correspondence definitions. The description of each rule will be explained with reference to FIG.

Rule 1701 is a rule for converting data in a database into a frame and storing it in the knowledge base. When the condition part 1702 of the rule 1701 is satisfied, the execution part 1703
execute the command. Execution unit 1703 has rea+1t
able1704 is an instruction to convert database data into a frame. The execution unit 1703 executes the above-mentioned “
Data in the database is converted into a frame based on the definition written in the frame "Conversion Definition 1". Rule 1705 is a rule for converting frames in the knowledge base into data and storing the data in the database. Rule 17
When the condition part 1706 of 05 is satisfied, the instruction of the execution part 1707 is executed. The execution unit 1707 has vriteJab
le 1708 is an instruction to convert the frame into database data. The execution unit 1707 converts the frame into database data based on the definition written in the frame "storage definition 1."

Rule 1709 is a rule for accessing database data. When the condition part 171O of the rule 1709 is satisfied, the execution part 1711 is executed. Delete 1712 held by the execution unit 1711 is an instruction to delete data in the database. The execution unit 1711 deletes data having the value "1" in the field corresponding to the slot "Company Capital" in the frame "Conversion Definition 1".

FIG. 18 shows the realized value of the class frame shown in FIG. 14, and is a part of the frame obtained by converting the data extracted from the table shown in FIG. Frame 1801 is field ``company name'' of table 201 in FIG.

This is a result of converting the data extracted from "employees", "capital", and the field "company name" of table 202. In the table 201, the value rA company' of the field r company name is held in the slot with the frame name '$Company name', the value r100J of the field 'Number of employees' is held in the slot 'Number of employees', and the field 'Capital gold” value “1”
is held in slot r company's capital.

Frame 1802 is also extracted from the same fields as frame 1801, and has the value "8 companies" in field r company name of table 201 in the slot "$company name" with the frame name, and has the field " The value r300J of "Number of Employees" is in the slot "Number of Employees", and the value "5" of the field "Capital" is set in the slot "Company Capital".
have it. Frame 1803 is the result of converting data extracted from fields ``Mr.'', ``1 Person'', and ``Special Skills'' of table 202. The value “Yamada” of the field “Mr.” in the table 202 is changed to the slot “$Last name” with the frame name.
, the value of the field "name" is set to the slot "$name" with the frame name, and the field "special skill" is
has the value [tennis, sailing, skiing] in the slot ``human special skills''.

Frame 1804 includes field “
This is the result of converting the data extracted from "Mr.", "1 Person", and "Company Name". The value ``Yamada'' of the field ``Mr.'' in table 202 is held in the slot ``$Last Name'' with the frame name, the value ``Afat'' of the field 1 name is held in the slot ``$Name'' with the frame name, and the field rCompany name”
The frame name of each frame that has the value "8 companies" in the slot "$company name" with the frame name differs depending on the contents of the conversion definition.

(1) When a frame is defined as an entity and there is only one field with a frame name, the value of that field is used as is for the name.

(2) When a frame is defined as an entity and there are multiple fields with frame names, the name is used by concatenating the values of the multiple fields.

(3) A frame defined as a relationship is used by concatenating the class name to which the frame belongs and the value of the field having the frame name.

An example of g(i) above is frames 1801 and 1802, an example of (2) is frame 1803, and an example of (3) is frame 1804.

Figures 19, 20, and 21 explain operation examples when converting database data to a frame, converting a frame to database data, and accessing database data from within a rule. It is.

FIG. 19 is an example of converting database data into a frame, and a frame 1803 is illustrated. Field “Mr.” of table 202.

"1 person" is a field containing a frame name.

It is assumed that tuples having the same value in this field represent one frame, and the value of the field "r special skill" that this tuple has is taken into one frame. In FIG. 19, the values of the fields "Mr.", "1 Name", and "Special Skill" of the tuples 1901, 1902 and 1903 represent the frame 1803.

FIG. 20 is an example of converting a frame into database data, and illustrates a table 202. The table 202 is a collection of r companies as can be seen from the relationship between the "collection name" 902 and the "table name" 903 in FIG.
, ``There is a relationship between the three sets of employee j and r work j. Therefore, the data of the table 202 is created from the relationship of the three sets.In other words, in Figure 9, the value r person'' is entered in the ``table name'' 904. Among the frames that belong to the class of "set name" 903 that has a certain value in "field name" 906 and has the value "1" in "key judgment" 912, the frame shown in "attribute name" 905 The value of the slot is “Collection name” 903 associated with the related entity I DJ 913
if it exists in a slot with the frame name of a frame belonging to the class .

Among these slots, the value of the slot that has a value in the "field name" 906 is stored in the database. In the example of FIG. 20, the slot value in frame 1802 is "r$company name," and the slot value in frame 1803 is "r$last name."

Set the value of the slot called "$Name" to "$Company Name".

Since frame 1804 exists in the slots "$Last Name" and "$Name," frame 1804 exists in the database.
The slot values of "$Company Name" in frame 02, "$Last Name" and "$Name", and "Special Skills of One Person" in frame 1803 are stored as a set. To store a frame, after deleting the data in the table, a tuple is created from the frame and stored in the table. Furthermore, if the slot of the frame to be stored has multiple values, such as frame 1803, one tuple is created for each value and stored in the table.

FIG. 21 shows an example of accessing database data from within a rule, and illustrates the table 201. In the inference engine 104, when the condition 1710 is satisfied and the rule 1709 is executed, rdeleteJ1712 of the execution unit 1711 is executed. rdelet
The eJ1712 deletes from the table the tuple that has the value "1" in the field corresponding to the slot "capital of r company" in the above-mentioned "conversion definition 1". The slot “Capital of r company in correspondence definition I J1301” in FIG. 13 is associated with the field “Capital” of the table 201.

In the table 201 of FIG. 21, there are two tuples having the value II I 11 in the field "capital".

When rdeleteJ1712 is executed, table 2
101, the value II in the field called "Capital"
There are no more tuples with I Tt. If a rule includes an instruction (search instruction) to retrieve a value from a table, a variable is prepared for that instruction to store the retrieved value. When a search command is executed, the data retrieved from the database is stored in the variable.

FIG. 22 illustrates the user interface 101. The user inputs to the user interface 101 by operating a text editor.

Includes operation of correspondence definition support system and expert system. A user input reading mechanism 2201 is a mechanism for reading those operating instructions, input data, and mouse point and ink information. Information read by the user input reading mechanism 2201 is sent to the processing selection mechanism 2202. The process selection mechanism 2202 determines the mechanism to be executed from the instructions contained in the information sent from the user input reading mechanism 2201, and sends data and instructions to other models. Also, from the information sent from the other model through the interface 2205 with other models, it determines the process to be executed and sends data and instructions to the other model.

In the case of translator execution, the translator 109
If the text editor is activated, the execution command is sent to the text editor 2203. For information display, data and commands are sent to display mechanism 2206. When executing a mechanism that the user interface 101 does not have, an execution command is sent to the interface 2205 with other models. The text editor 2203 also uses the processing selection mechanism 2.
It is operated by a command from 202 and sends a text display command to processing selection mechanism 2202 . A text editor 2203 edits text information in a text space 2204 that the user interface 101 has. Also, the text in the text space is stored in the file system 111. The interface 2205 with other models determines the command from the process selection mechanism 2202 and connects the inference engine 104 . Correspondence definition support system 103. Send commands to database-expert system coupling interface 106. Inference engine 104. It receives data and instructions from the correspondence definition support system 103 and sends the information to the processing selection mechanism 2202.

FIG. 23 illustrates the correspondence definition support system 103. The input from the user interface 101 is
The processing selection mechanism 2301 is entered. The processing selection mechanism 2301 determines the command from the user interface 101 and sends the command to other mechanisms. When a database search command comes from the user interface lot, the process selection mechanism 23
01 sends a command to the database search mechanism 2302.

When input of a set or an attribute of a set is received from the user interface 101.

A command is sent to the diagram determination processing mechanism 2303. When a command to divide or integrate a set comes from the user interface 101, the command is sent to the command processing mechanism 2304.

When a command to end correspondence definition support is received from the user interface 101, the command is sent to the correspondence definition creation machine MIt2305. The database search mechanism 2302 sends a database search command to the database management system 107 in response to an instruction from the process selection mechanism 2301 to perform a database search.

The database search mechanism 2302 receives the searched data and sends the search data and data display command to the user interface execution function 2307. The diagram determination processing mechanism 2303 refers to the internal table 2306 (see FIG. 9) in accordance with the command from the process selection mechanism 2301.

Perform updates. When the setting of the internal table 2306 is completed, an instruction and data for displaying a model representing the state of the internal table 2306 at that time are sent to the user interface execution function 2307.

The command processing mechanism 2304 refers to the internal table 2306 and updates the table to a divided or integrated state based on the contents of the table. When the update of the internal table 2306 is completed, an instruction and data for displaying a model representing the state of the internal table 2306 at that time are sent to the user interface execution function 2307. The correspondence definition creation mechanism 2305 selects the first
Generate the correspondence definition in Figure 3 and the class frame in Figure 14,
The instruction to start the text editor, the instruction to store text in the text space of the user interface 101, and the created data are sent to the user interface execution function 2.
Send to 307.

The user interface execution function 2307 performs processing for passing commands and data from other models to the user interface 101.

FIG. 24 illustrates the inference engine 104.

Inference engine 104 operates according to instructions from user interface 101 . Knowledge base interpretation mechanism 240
1 reads frames and rules from the knowledge base 105 in response to an activation command from the user interface 101, interprets the contents of the frames and rules, and sends the interpreted results to the execution rule selection mechanism 2402. When a termination command is received from the user interface 101, the termination command is sent to the execution rule selection mechanism 2402. Execution rule selection mechanism 2
402 selects one rule to be executed from the interpretation results sent from the knowledge base interpretation mechanism 2401 and sends the execution part of that rule to the process determination 2403 . Inference ends when there are no more executable rules. When a termination command is sent from the knowledge base interpretation mechanism 2401, no rule is selected.

The process determination 2403 determines whether there is a process using a database in the execution part of the rule sent from the execution rule selection mechanism 2402, and if a database is used, the command is sent to the database-expert system combination interface 106. send.

If the process does not use a database, run rule execution 2
Send the command to 404. Rule execution 2404 is a mechanism that actually executes the execution part of a rule, and when an output command to user interface 101 is executed, it sends an output command to the user interface.

Furthermore, when a write command is executed to the knowledge base 105,
Write to the knowledge base 105. Rule execution 2404
sends the executed content to the execution rule selection mechanism 2402 and uses it to select the next rule. Information from the database-expert system coupling interface 106 is also used to select the next rule.

FIG. 25 is an explanatory diagram of the database-expert system coupling interface 106. Commands from the user interface 101 are sent to the processing determination mechanism 2501. When the processing determination mechanism 2501 receives an instruction from the user interface 101 to convert database data into a frame and store it in the file system 111, it sends the instruction to the search command creation mechanism 2502. The search command creation mechanism 2502 reads the correspondence definition from the file system 111 through the frame translator 108, creates a database search command, and sends the search command to the database management system 107.

This causes the data frame modification mechanism 2506 to retrieve data from the database management system 107, convert it into a frame, and send it to the frame translator 108.
The data is stored in the file system 111 through.

Further, when a command is sent from the user interface 101 to convert a frame in the file system 111 into database data and store it in the database, the command is sent to the update command creation mechanism 2505. The update command creation mechanism 2505 reads the correspondence definition from the file system 111 through the frame translator 108,
Create an update command. As a result, the frame → data conversion mechanism 2507 reads the frame from the file system 111 through the frame translator 108,
Convert it to data and store it in the database.

Instructions from the inference engine 104 are processed by a processing determination mechanism (I'
E) It is sent to 2504. When the processing determination mechanism 2504 receives a command from the inference engine 104 to convert database data into a frame and store it in the knowledge base 105, it sends the command to the search command creation mechanism 25o2. The search command creation mechanism 2502 uses the knowledge base 10
The correspondence definition is read from 5, a database search command is created, and the search command is sent to the database management system 107. As a result, the data frame modification mechanism 2506 changes the database management system 1
07, converts it into a frame, and stores it in the knowledge base 105. Further, when the inference engine 104 receives a command from the user interface 101 to convert frames in the knowledge base 105 into database data and store the same in the database, the inference engine 104 sends the command to the update command creation mechanism 2505 . Update command creation mechanism 25
05 reads the correspondence definition from the knowledge base 105 and creates an update command. Thereby, the frame→data conversion mechanism 2507 reads the frame from the knowledge base 105, converts it into data, and stores it in the database.

When an instruction to access the database is sent from the inference engine 104, the direct access mechanism 250
Send command to 3. The direct access mechanism 2503 is
In response to an access command from the inference engine 104, the search command creation mechanism 2502. Update command creation mechanism 25
Send command to 05. Search command creation mechanism 2502. The update command creation mechanism 25o5 each generates a command and sends the command to the database management system 107. In addition, the direct access mechanism 2503
If the access command from the inference engine 104 is a search command, sends the data retrieved from the database to the inference engine 104.

In this embodiment, the expert system and database system are all combined in one subsystem and can be established as a subsystem, so if there is no need to combine the database and expert system,
Subsystems can also be removed. This has the effect that a system can be selected depending on the performance of the computer system or how the expert system is used.

Another embodiment of the present invention will be shown below. Figure 26 is.

The database-expert system combination interface 106 described in FIG. 25 is divided into two parts: a database-expert system combination interface 2601 compatible with a user interface, and a database-expert system combination interface 2602 compatible with an inference engine. In FIG. 25, the user interface 101 being sent to one subsystem
Instructions from the user interface compatible database-expert system coupling interface 2601
Sent to.

Instructions from the inference engine 104 are sent to the inference engine compatible database-expert system coupling interface 2602. This makes it possible to select subsystems more precisely than in the embodiments described above, and to simplify the processing of each mechanism.

Next, still another embodiment of the present invention will be described using FIG. 27. FIG. 27 shows the database-expert system coupling interface 10 explained in FIG.
6, a first database-expert system coupling interface (managing data frame conversion) 2701 compatible with a user interface, and a second database-expert system coupling interface (managing frame-data conversion) 2701 compatible with a user interface.
702. First database-expert system coupling interface compatible with inference engine (controls database access) 2703. Inference engine compatible second database-expert system coupling ink interface (controls frame-data conversion) 2704; inference engine compatible third database-expert system coupling interface (controls data frame conversion) 2705 It is divided into In FIG. 25, the commands from the user interface 101 that were sent to one subsystem are sent to the two database-expert system coupling interfaces 27 mentioned above.
01.2702 and instructions from the inference engine 104 are sent to the three database-expert system coupling interfaces 2703.2704.2705 mentioned above. As a result, it is possible to select subsystems more precisely than in the second embodiment described above, and the processing of each mechanism can be made simpler than in the above embodiment.

Next, still another embodiment of the present invention will be described using FIGS. 28, 29, and 30.

FIG. 28 shows a display screen of the user interface 101. Screen 1001 is a table list display screen, screen 1
002 is the database search result display screen 1 screen 100
3 is a modeling display screen, and screen 1601 is an execution screen of the expert system. The user can switch screens by pointing at these screens with the mouse pointer 1008. In this way, two systems can be operated simultaneously and multiple screens can be displayed simultaneously.

This can be achieved by using a user interface 2901 as shown in FIG. 29 in place of the user interface 101, and by using a correspondence definition support system 3001 in FIG. 30 in place of the correspondence definition support system 103.

A window switching mechanism 2902 included in the user interface 2901 determines that the user has switched the window with the mouse pointer 1008, and sends a command to the task management mechanism 2903 when the window is switched.

The task management mechanism 2903 determines which screen the mouse pointer 1008 is pointing to, and when a task different from the currently running task is selected, sends a command to the process selection mechanism 2905 to temporarily suspend the original task.

For example, when the expert system execution screen 1601 is pointed to while the correspondence definition support system 3001 is being executed.

The correspondence definition support system 3001 is temporarily suspended and the expert system is executed. In addition, the task management mechanism 29
03 sends a command to the window synchronization mechanism 2904 to make the selected screen the topmost window.

The window synchronization mechanism 2904 makes the selected screen the topmost window.

The window switching mechanism 3002 of the correspondence definition support system 3001 determines that the user has switched windows with the mouse pointer 1008, and when the windows are switched, causes the window synchronization mechanism 3003 to make the selected screen the topmost window. . In this embodiment, the user can view two systems at the same time, so the user can easily use the knowledge processing system.

〔Effect of the invention〕

According to the present invention, a user can easily understand the correspondence between database data and frames using a correspondence definition system that allows correspondence between database schemas and class frames on the screen. In addition, the correspondence definition system that can send the correspondence definition results to the editor allows the user to add and modify the necessary information to the correspondence definition results, store it in a file, and use this information as is in the expert system.It is a knowledge processing system. , users can easily construct processes using databases. From the user interface, you can convert database data into frames and store them in a file, or from the inference engine you can convert database data into frames and store them in a knowledge base. The frames can be stored in a database, and the database can be accessed directly from the inference engine, making it possible to use it in a variety of ways.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the configuration of the knowledge processing system, FIG. 2 is a diagram showing an example of a database used in the embodiment, FIG. 3 is a display of an entity-relationship set, and FIG. 4 is an execution flow of the knowledge processing system. Figure 5 is the table selection process flow, Figure 6 is the correspondence definition process flow, Figure 7 is the model division process flow, and Figure 8 is the table selection process flow.
The figure shows the model integration processing flow, Figure 9 shows the internal table, Figure 10 shows the display screen of the correspondence definition support system, Figure 11 shows
Figure 12 is a diagram showing the display state of the model, Figure 13 is a diagram showing an example of correspondence definition, Figure 14 is a class frame determined by the correspondence definition, Figure 15 is a correspondence definition class frame,
Figure 17 shows the execution screen of the expert system, and Figure 17 shows an example of rules using a database. Figure 18 shows an example of an instance frame, Figure 19 shows a method of converting data in a table into a frame, Figure 20 shows a method of storing frames in a table, Figure 21 shows the operation of a database access command, and Figure 22 shows a method of converting data in a table into a frame. 23 is the processing configuration of the correspondence definition support system of the embodiment, FIG. 24 is the processing configuration of the inference engine of the embodiment, and FIG. 25 is the database-expert system combination of the embodiment. Processing configuration of the interface, Fig. 26 shows the processing configuration of the database-expert system combination interface of the embodiment, Fig. 27 shows the processing configuration of the database-expert system combination interface of the embodiment, and Fig. 28 shows the processing configuration of the knowledge processing system of the embodiment. The display screen, FIG. 29 shows the processing configuration of the user interface of the embodiment, and FIG. 30 shows the processing structure of the correspondence definition support system of the embodiment. 101: User interface, 102: User, 10
3: Correspondence definition support system, 104: Inference engine, 1
05: knowledge base, 106: database-expert system combination interface, 107: database management system, 108: frame translator, 109:)-translator, 11O = database system, 111: file system. 1001: Table view, 1002: Database search screen, 1003: Model display screen, 1301:
Conversion definition 1.1314: Storage definition 1.1601: Expert system execution screen, 2201: User input reading mechanism, 2202: Process selection mechanism, 2203: Text editor, 2205: Interface with other mechanisms, 2206: Display mechanism, 2301: processing selection mechanism;
2302: Database search mechanism, 2303: Diagram determination processing mechanism, 2304: Command processing mechanism, 230
5: Correspondence definition creation mechanism, 2306: Internal table, 2
307: User interface execution mechanism, 2401:
Knowledge base interpretation mechanism, execution rule selection mechanism, 2403
: Processing determination, 2404 : Rule execution, 2501 : Processing determination mechanism, 2502 : Search command creation mechanism, 25
03: Direct access mechanism, 2504: Processing determination mechanism. 2505: Update command creation mechanism, 2506: Data reframe conversion mechanism, 2507: Frame → data conversion mechanism, 2902: Window switching mechanism, 2903
: Task management mechanism, 2904 : Window synchronization mechanism. Figure Command: 1--Ditor:--→〉 Command and Data Knee-U5 Figure Figure Start Figure (Part 1) Figure (Part 2) Figure (Part 1) (Initial Display (2) Entity (14061) members. ,. r-cjass1407
□$ Last name 1408 Shiva 8 Name 1409 (7-person's special skill system trlng str+ng strlng 5et 1411, $Company name 1412 (/-$Last name 1413 (7-8 Name 1 Dumpling 1 Corridor 3 further trtng 17th Figure 1701 ~ (Table reading rule 1702 ~ if ('i'X@c/ass conversion definition frame) 1703 ~ then (send conversion definition 1 read-table) 17
05~(Table write rule 1706~if ('i'X @class storage definition frame) 1707~then (send storage definition l write-1able) 1
709VS (Table access rule 171QI-/-
if ('i'X@c/ass conversion definition frame) 1
711 vthen (send conversion definition 1 delete (company capital,
1)) Figure V (Company A class Company $ Company name Number of employees at Company A 100 Figure $ Company name Number of employees Company B $ Last name 8 Name Yamada Ata Company name $ Last name 8 names Company B Yamada A) $ Last name 8 Famous skills Ata Yamada (tennis, sailing, skiing)) 1802 ', - (Company B Figure 1ass $ Name of the company Number of employees Company capital 18o3 (... Ata Yamada ' 1ass 8 Last name 8 Special Skills Company B Employee Yamada Requests Tennis) No.21
if ('i'X @class1711y"t
hen Conversion constant a frame) $ Company name $ Last name 8 Name Company B Yamada Ata

Claims (1)

[Claims] 1. Correspondence between relational database data constructed and managed in a specific structure on secondary storage and the frame of an expert system that constructs a knowledge base on main memory and performs inference using an inference engine. The method includes the steps of selecting a table that a user needs in an expert system from tables held in the relational database; a step of determining an attribute having a frame name; a step of determining which attribute information the information of an attribute without a frame name belongs to; and a step of dividing the entity set into a plurality of entities using these pieces of information. The step of creating entity sets and relation sets that connect them is performed for all the tables selected by the user, and by integrating sets with the same name, a model using entity sets and relation sets is created, and each set is A method for supporting definition of correspondence between an expert system and a database system, characterized by representing it as a class frame. 2. A knowledge processing system using the method for supporting correspondence definition between an expert system and a database system according to claim 1, wherein the knowledge processing system uses an interface between the user and the database data and the expert system based on the entity set and the relationship set. an expert system having a correspondence definition support system for associating frames, a knowledge base that stores frames and rules, and an inference engine that performs inference using the frames and rules in the knowledge base; A database management system that performs management, a joint interface between a database and an expert system that performs interface processing between the database and the expert system via the database management system, and a format that allows frames in a file system to be stored in a knowledge base. The present invention is characterized by comprising a frame translator that converts the data into a text format, sends it to a database-expert system combination interface, converts it from a form that can be stored in a knowledge base to a text format, and stores it in a file system, a database, and a file system that stores the file. knowledge processing system. 3. The conversion definition support system receives a command from the user interface, and includes a means for selecting a support process to be performed by the support system, and a means for identifying a chart on the screen from the screen information sent from the user interface to display the screen. A means for determining the information in the above chart and outputting the result, and a means for converting the information held in the internal table and outputting the result when a command for dividing or integrating entity-relationship sets is input from the user interface. means for sending a data search command to the database management system when a database search command is input from the user interface, receiving the search results and outputting the results; and means for outputting the results from the support system. 3. A knowledge processing system according to claim 2, further comprising means for sending information to said user interface. 4. The user interface includes a means for reading user input from a keyboard or a mouse as a user input means, a means for displaying diagrams and text on a display as an output means for the user, and a knowledge processing system based on information input by the user. 3. The knowledge processing system according to claim 2, further comprising means for selecting a process to be performed, and a text editor for creating and editing text and storing the results in a file system within the expert system. 5. The inference engine includes a means for interpreting information in the knowledge base, a means for selecting a rule to be executed using the interpreted knowledge base information and the execution result of the rule, and a process of the rule to be executed. A means for determining whether to use a database, sending a command to the database-expert system coupling interface when using a database, and executing a rule when not using a database, and a means for executing a rule based on the result of executing the rule. 3. The knowledge processing system according to claim 2, further comprising means for selecting power rules. 6. The database-expert system coupling interface receives input from the user interface;
Input/output with the inference engine, input/output with the knowledge base, input/output with the frame translator,
means having an input/output with the database management system and determining a process to be executed from information input from a user interface or an inference engine;
When receiving a command from the inference engine to convert the data in the database into a frame and store it in the knowledge base, the correspondence definition between the corresponding database and expert system is retrieved from the knowledge base, a database search command is generated from this, and the data is stored in the database. a means for sending search results to a management system, a means for receiving search results and converting them into frames with unique names and storing them in a knowledge base;
When receiving a command from the inference engine to convert a frame in the knowledge base to data in the database and store it in the database, it retrieves the correspondence definition between the corresponding database and expert system from the knowledge base, and executes the database data storage command from this. means for generating and sending it to a database management system; means for converting frames extracted from the knowledge base into data and storing them in a database; and generating database access commands from the inference engine using the correspondence definition between the data in the database and the expert system. means for selecting a process according to the access command when received, sending search results to an inference engine if it is a search command, updating the database if it is an update command, and converting data in the database into a frame from the user interface. When receiving an instruction to store the search result in the file system, a means for extracting the correspondence definition between the corresponding database and the expert system from the file system, creating a database search command from this and sending it to the database management system, and receiving the search results. means for converting the frame into a frame with a unique name and storing it in the file system; A means for extracting a correspondence definition between a database and an expert system, creating a database data storage command from this and sending it to a database management system, and a means for converting a frame in a file system to data in a database and storing it in the database. The knowledge processing system according to claim 2, characterized in that: 7. The means for identifying the chart on the screen from the screen information sent from the user interface and outputting the information on the chart on the screen uses a multi-window,
means for determining that the user has switched windows;
Displaying the database table framework accordingly,
4. The knowledge processing system according to claim 3, further comprising window synchronization means for simultaneously displaying database search results and entity-relationship sets using the multi-windows. 8. The database-expert system coupling interface receives input from the user interface;
has an input/output with the frame translator and an input/output with the database management system,
a means for determining a process to be executed from information input from the user interface; and a means for determining a process to be executed from information input from the user interface, and when receiving an instruction from the user interface to convert data in a database into a frame and storing it in a file system, converting the corresponding database from the file system. a means for extracting the corresponding definition of the expert system and generating a database search command from this and sending it to the database management system; a means for receiving the search result and converting it into a frame with a unique name and storing it in a file system; When receiving a command from the interface to convert a frame in the file system to data in a database and store it in the database, it retrieves the correspondence definition between the corresponding database and expert system from the file system, and generates a data storage command for the database from this. and a means for converting a frame retrieved from a file system into data and storing it in a database, input/output between the inference engine, and a knowledge base. a means for determining a process to be executed from information input from the inference engine, and a means for determining a correspondence between database data and the expert system from the inference engine; means for selecting a process according to the access command when receiving a used database access command, sending search results to an inference engine if it is a search command, updating the database if it is an update command; When receiving an instruction to convert the data into a frame and store it in the knowledge base in the expert system, extract the correspondence definition between the corresponding database and the expert system from the knowledge base,
A means for creating a database search command and sending it to the database management system, a means for receiving the search results and converting them into frames with unique names and storing them in the knowledge base, and a means for converting the frames in the knowledge base from the inference engine to the database. means for extracting a correspondence definition between the corresponding database and the expert system from the knowledge base, creating a data storage command for the database from the knowledge base, and sending the command to the database management system when receiving an instruction to convert the data into a database and store it in the database; 7. The knowledge processing system according to claim 6, wherein the knowledge processing system is divided into a second part having means for converting frames in the knowledge base into data and storing the data in the database. 9. The database-expert system coupling interface receives input from the user interface;
has an input/output with the frame translator and an input/output with the database management system,
When receiving a command from the user interface to convert the data in the database into a frame and store it in the file system, the correspondence definition between the corresponding database and expert system is retrieved from the file system, a database search command is generated from this, and the data is stored in the database. a first part having means for sending the search results to a management system and means for receiving and converting the search results into uniquely named frames and storing them in a file system; input from the user interface; input/output, input/output to and from the database management system, and when receiving an instruction from the user interface to convert a frame in the file system into data in the database and store it in the database, A second device having means for extracting the correspondence definition between the database and the expert system, generating a database data storage command from this and sending it to the database management system, and means for converting the frame extracted from the file system into data and storing it in the database. , an input/output between the inference engine, an input/output with the knowledge base, and an input/output with the database management system, and processing to be executed from information input from the inference engine. and when receiving a database access command from the inference engine using the correspondence definition between database data and expert system, selects a process according to the access command, and if it is a search command, the inference engine sends the search result to the inference engine. and input/output between the inference engine, the knowledge base, and the database management system. and when it receives an instruction from the inference engine to convert data in the database into a frame and store it in the knowledge base in the expert system, it extracts the correspondence definition between the corresponding database and the expert system from the knowledge base, and extracts it from the knowledge base. a fourth part having means for creating a database search command and sending it to the database management system; and means for receiving and converting the search results into a frame having a unique name and storing the frame in the knowledge base; and the inference engine. has an input/output between the interface, an input/output with the knowledge base, and an input/output with the database management system, and converts frames in the knowledge base from the inference engine into data in the database and stores it in the database. When a command is received, a means for extracting the correspondence definition between the corresponding database and expert system from the knowledge base, creating a database data storage command from this and sending it to the database management system, and converting the frames in the knowledge base into data. 7. The knowledge processing system according to claim 6, wherein the knowledge processing system is divided into a fifth part having means for storing in a database.