JPH0991108A - Interactive operation supporting device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To receive work support even without applying any exact instruction by inputting the approximate purpose of work from an input/output means corresponding to several keywords by a user and conforming to guidance based on interactive flow knowledge. SOLUTION: An interactive managing part 1 interpretes the inputted keyword while utilizing a tag DB 5 and a special tag DB 6 provided at a knowledge managing part 2. The interpreted result is recorded in a topic managing table 7 and a parameter managing table 8 and the interactive flow knowledge is executed successively from the knowledge having the highest execution priority recorded in the topic managing table 7. The inquiry to the user, the execution of application program and condition judgement using provided data or the like are described in the interactive flow knowledge and the interactive managing part 1 executes the interactive flow knowledge while utilizing data recorded in the parameter managing table 8. The user can input the keyword any time and can freely execute the interactive flow by switching it.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an interactive information processing apparatus capable of interacting with a user, and more particularly to an interactive operation supporting apparatus capable of assisting a user operation while interacting.

[0002]

2. Description of the Related Art In a conventional information processing apparatus, it is premised that a user knows a correct method of use, and when operating an information processing apparatus, a command and its parameters are correctly combined and in a correct order. Had to enter. As an invention for solving such a problem, the following technology relating to an interactive information processing apparatus has been disclosed.

Japanese Patent Application Laid-Open No. 02-202613 discloses a "user interface management device for an interactive information processing apparatus".
A command database is used that stores a storage unit, which is a set of a command identifier, a parameter condition required by a command, and three pieces of information of a menu, for all commands. The command is limited by checking whether the user's input matches parameter conditions and menu items.

On the other hand, Japanese Patent Laid-Open Publication No. Hei 7-56853, entitled "Dialog Support Knowledge Representation Method", defines dialog flow knowledge, which is knowledge for interactively supporting a user to use the system. Japanese Patent Application Laid-Open No. 7-192052 discloses an information management method for enabling a user to reuse information obtained in a system that interactively supports a user's use of the system by using the interaction flow knowledge. A method is defined.

[0005]

However, in Japanese Patent Application Laid-Open No. 02-202613 described above, the user must know the meaning and action of each command, and also to achieve the work purpose. However, there is a problem that the user has to consider a combination of commands.

In Japanese Patent Application Laid-Open Nos. Hei 7-56853 and Hei 7-192052, which are methods using dialog flow knowledge,
There is a problem that it becomes difficult for the user to select when the dialog flow knowledge increases, and it is not possible to respond to the change of the free topic according to the input of the user,
There has been a problem that no method has been considered for correcting the obtained information.

[0007] An object of the present invention is to solve the above-mentioned problems of the prior art by allowing a user to input an approximate work purpose from input / output means by using some keywords and to follow guidance based on dialog flow knowledge. By doing
An object of the present invention is to provide an interactive operation support device that can receive work support without giving an accurate instruction.

[0008]

The first invention of the present invention is as follows:
In an interactive system operation support device that supports the execution of work by giving appropriate guidance to users who work by combining multiple application programs or the commands of application programs, the user's work purpose is keyworded Input and output means for presenting the obtained data to the user, guidance such as questions and explanations of operation methods, explanation of questions and operation methods to the user, execution of application programs, Flow knowledge base that holds a plurality of dialog flow knowledge in which condition judgments and the like are described using the extracted data, a tag DB in which parameter names and their values are extracted from the dialog flow knowledge base, and a specific A special tag DB that extracts tags used only in the dialog flow knowledge, A knowledge management unit that generates a tag DB from the talk flow knowledge base, generates a special tag DB from the tag DB, and outputs the contents of the dialog flow knowledge base, the tag DB, and the special tag DB according to a request from the dialog management unit; A knowledge execution unit that actually starts an application program or transmits a command sequence to the application program in accordance with an instruction, and a dialog flow to be executed based on a result of a keyword input interpretation from a user or an execution result in the knowledge execution unit. An information management unit that manages using a topic management table that records knowledge, and a parameter management table that stores obtained data as a set of an attribute name and an attribute value; and the tag DB via the knowledge management unit. The special tag DB is used to interpret the keyword input by the user, and the interpretation result is transmitted to the information management unit and recorded. A dialog management unit for giving an instruction to the knowledge execution unit using data recorded by the information management unit according to the description of the plurality of dialog flow knowledge of the dialog flow knowledge base via the knowledge management unit. Features.

According to a second aspect of the present invention, in the first aspect, the tag DB stores the dialog flow name of the dialog flow knowledge, a parameter name used in the dialog flow knowledge, and its value. Tag DB extracted as
The special tag DB is a special tag DB that extracts tags used only for specific dialog flow knowledge from the tag DB, and includes a keyword input by a user and the tag DB.
By comparing the tag with the tag extracted in the special tag DB, interpretation of a user input, change of a topic, generation of a question to the user, and the like are performed.

Further, according to a third aspect of the present invention, in the first or second aspect, the topic management table includes an execution number indicating a topic execution priority, a dialog flow name, and a parameter management table corresponding to the dialog flow. Number, the node name indicating the part of the dialog flow that has been executed,
A topic management table having a field such as a parent topic indicating whether it has been called from another conversation flow, and using the topic management table, recording a result of user's input interpretation and an execution result in the knowledge execution unit. Thus, the execution order of the dialog flow can be freely changed.

Further, the fourth invention of the present invention provides the first and second inventions.
Alternatively, in the third invention, the parameter management table includes:
A parameter management table having a parent data field for recording data used as a search key for acquiring the data other than the attribute name and the attribute value of the information obtained by the dialogue, wherein the parameter management table is used. By
When the data is modified, only data related to the data is extracted.

[0012]

Embodiments of the present invention will now be described with reference to the drawings.

Some keywords input by the user through the input / output means 9 are passed to the dialogue management unit 1.
The dialogue management unit 1 uses the tag DB 5 and the special tag DB 6 possessed by the knowledge management unit 2 to interpret the input keyword,
The information management unit 3 is instructed to record the interpretation result in the topic management table 7 and the parameter management table 8. The tag DB 5 and the special tag DB 6 are automatically updated by the knowledge management unit 2 when the dialog flow knowledge base 4 is updated.

If a dialog flow name that is not written as “end” is recorded in the node name field of the topic management table 7, the dialog management unit 1 instructs the knowledge management unit 2 to execute the execution priority with the highest execution priority. The high dialog flow knowledge is extracted from the dialog flow knowledge base 4. Dialog flow knowledge includes
Questions to users, execution of application programs,
A condition judgment or the like using the obtained data is described, and the dialog management unit 1 executes the dialog flow knowledge by using the data recorded in the parameter management table. A question is asked to the user through the input / output unit 9 and the application program is executed through the knowledge execution unit. The obtained data is recorded in the parameter management table 8 and used for executing the dialog flow. Even during the execution of the dialog flow, the user can input a keyword through the input / output means at any time, and the dialog management section 1 interprets the input keyword and instructs the information management section 3 to instruct the topic management table 7. Can be freely switched to execute the interactive flow.

If a call of another dialog flow is described in the dialog flow, the name of the called dialog flow is added to the topic management table 7, and the parent topic field of the topic management table 7 is added. Write the name of the called conversation flow to Depending on the user's input, the order of the dialog flows to be executed may be switched, and the calling dialog flow may be executed before the called dialog flow. Even in such a case, it is checked whether the called conversation flow name is recorded in the topic management table 7, and if it is recorded, whether the called conversation flow name is written in the parent topic field. Therefore, the execution order can be corrected so that the called conversation flow is executed first.

When the data obtained during execution is modified, the parent data field of the modified data in the parameter management table 8 is checked. In the parent data field, data used as a search key to acquire the data is recorded, so check data having the same parent data as the data to be corrected, check the parent data of the checked data, Repeat recursively, checking for data with the same parent data. When the check is completed, the data with the check is deleted and the specified data is corrected.

Thereafter, the dialog flow using the parameter management table 8 in which the corrected data is recorded is searched by checking the table number field of the topic management table 7, and the node name field of the searched dialog flow is searched. To "Start".

Although the dialog flow related to the corrected data is re-executed, the data already recorded in the parameter management table is not re-executed or the user is not asked the same question repeatedly. Thus, only necessary data can be efficiently acquired.

Next, a more detailed embodiment of the present invention will be described with reference to the drawings and examples.

FIG. 1 is a block diagram showing an embodiment of the configuration of the interactive operation support apparatus according to the present invention.

The knowledge management unit 2 manages a dialog flow knowledge base 4 described in a flow format. For each dialog flow knowledge in the dialog flow knowledge base 4, a dialog flow name and a description described in the knowledge are provided. Then, the parameter name and its value used for the conditional branch at the time of execution are registered in the tag DB 5 as a tag attached to the dialog flow knowledge (20).
5). However, if a call for another dialog flow knowledge is described in a certain dialog flow knowledge, the name of the called dialog flow is not registered as a tag of the called dialog flow knowledge. In addition, among all the dialog flow knowledge, a tag included only in a certain percentage (for example, 50%) or less of the dialog flow knowledge is checked (502) and registered as a special tag in the special tag DB 6 (206).

For example, as shown in FIG.
It has a dialog flow name of "power failure" and "turn off electricity".
Consider the case where there are three dialog flow knowledges.

With respect to the dialog flow knowledge of "turn on electricity", "turn on electricity", "payment not yet received", "moving"
For the dialog flow knowledge "power failure", "power failure"
"Blackout area""Customersonly""Wholearea""Construction"
However, with respect to the dialog flow knowledge of "stop electricity", "stop electricity", "moving", "contract type", "low-voltage power", "midnight power", and "farm power" are tag DB5 as tags, respectively. Is registered (205).

Further, it is assumed that the dialog flow knowledge is these three, and as an example, it is included only in the dialog flow knowledge of 50% or less (in this case, three dialog flow knowledges and 50% or less, one). Use keywords as special tags. In this example, "turn on electricity", "payments not yet received", "blackout", "blackout range", "customers only", "whole area", "construction", "stop electricity", "contract type", "low voltage power", and "midnight power "Agricultural power" is registered in the special tag DB 6 as a special tag (206). FIG. 3 shows the tag DB generated from the dialog flow knowledge of FIG. 2, and FIG. 4 shows the special tag DB.

The tag DB 5 and the special tag DB 6 are automatically generated from the dialog flow knowledge base by the knowledge management unit 2, but the user can add a new keyword after the generation.

The user expresses an approximate work purpose by using some keywords, and transmits it to the dialog management section 1 through the input / output means 9 (901). Regarding the input / output means 9, input from a keyboard, voice input / output, GU
Menu selection, button click, pen input, touch sensor, and the like using I (graphical user interface) can be used.

When the user inputs a keyword, the dialog manager 1 instructs the knowledge manager 2 (102, 201),
The input keyword is compared with the special tag registered in the special tag DB 6 (602). If the input keyword is included in the special tag, the dialogue management unit instructs the information management unit 3 to write all dialogue flow names including the special tag in the topic management table 7 (10
3, 307).

The information management unit 3 manages data using the topic management table 7 and the parameter management table 8. The topic management table 7 includes, as shown in FIGS. 5 and 7, the execution priority of the dialog flow knowledge, the dialog flow name, the node name indicating up to which part of the dialog flow knowledge has been executed, and the number of the corresponding parameter management table. It has fields such as a table number to represent, and a parent topic to indicate whether it was called from other dialog flow knowledge. The topic management table 7 is used to manage the priority of the dialog flow knowledge to be executed, and the dialog flow knowledge is executed in an order according to the execution priority. When the execution of the dialog flow knowledge is newly started, the execution priority of the dialog flow knowledge is set to the highest in the topic management table 7, and the dialog flow name and the number of the corresponding parameter management table 8 are written. "Start" is written in the node name field, and the execution is started (FIG. 5). The topic management table 7 is changed by the dialog management unit 1 interpreting the input keyword of the user and instructing the information management unit 3 (103) (30).
7), the execution priority of the dialog flow is changed.

The parameter management table 8 is composed of a plurality of tables as shown in FIG. 8, and each table has a unique table number. Although only one parameter management table is used at first, the number of used parameters increases each time new dialog flow knowledge is executed. Each table has, in addition to fields for recording parameter names and their values, a parent data field for recording data used as a search key to acquire the data.

For example, let us consider a case where the user inputs “payment not yet received” as a keyword from the input / output unit 9. First, the keyword "payment not yet received" is transmitted from the input / output means 9 to the dialog management section 1 (901). The dialog managing unit 1 instructs the knowledge managing unit 2 (102) and checks the special tag DB 6 (206, 602).

Here, it is found that the "payment due" is a special tag (FIG. 4) and is a tag attached to the dialogue flow "Turn on electricity" (201).
Instructs the information management unit 3 to write the conversation flow name "Turn on electricity" in the topic management table 7 (307) (103). When the dialog flow name is entered in the topic management table 7, one parameter management table 8 used when executing the dialog flow is prepared, and the table number of the parameter management table 8 is written in the topic management table 7 (80
3, 307). When a new dialog flow name is written in the topic management table 7, "start" is written in the node name field (FIG. 5).

The dialog managing unit 1 refers to the tag DB 5 via the knowledge managing unit 2 and finds that the input keyword is a tag other than the dialog flow name (102, 205, 5).
02, 201), if the group of the parameter name of the input keyword and the corresponding parameter value is known, the dialog management unit 1 instructs the information management unit 3 (103), and the parameter name and the parameter value are managed by the parameter management. Record in Table 8 (308). In addition, the dialog management unit 1 sets “YE
If only the parameter name of a parameter that takes two values of "S" and "NO" is input, it is interpreted that "YES" is specified as the value. In this example, the keyword “payment receivable” is a parameter name, and it can be seen from the tag DB5 that the value takes two values, “YES” and “NO” (102, 502, 201). In the parameter management table 8, "parameter name: unpaid fee, value: YE
S ”is written (103, 308). Nothing is written in the parent data field for the parameters input by the user (FIG. 6).

As described above, even when the user does not directly input the dialog flow name, the dialog flow name can be written in the topic management table, and the user's work support can be started according to the dialog flow knowledge. Can be.

Next, the dialogue management unit 1 instructs the information management unit 3 (103) to read out the dialogue flow name and the node name having the highest execution priority from the topic management table 7 (70).
3, 301).

If the topic management table is as shown in FIG. 5, the dialog flow name "turn on electricity" and the node name "start" are read.

The dialog management unit 1 instructs the knowledge management unit 2 (102) to read the dialog flow knowledge of the dialog flow name "turn on electricity" from the dialog flow knowledge base 4 (40).
2, 201). The dialog flow knowledge is executed continuously from the node name "start" recorded in the topic management table 7. The node name field is rewritten as the dialog manager 1 proceeds with the execution of the dialog flow (103, 3).
07).

In the dialog flow knowledge, a question to the user, execution of an application program, condition judgment using obtained data, and the like are described. A question is asked to the user through the input / output unit 9 (109, 901), and the application program is executed through the knowledge execution unit 10 (1010, 1001). The obtained data is recorded in the parameter management table 7, and is used by the dialog management unit 1 to execute the dialog flow knowledge.

Next, let us consider a case where the user newly inputs another keyword through the input / output means 9 during the execution of the dialog flow knowledge. The dialogue management unit 1 checks the keyword (901) passed from the input / output unit 9 using the tag DB5 and the special tag DB6 managed by the knowledge management unit 2 (5).
02, 602). If the input keyword is a parameter name or its value, the information management unit 3 is instructed to record it in the parameter management table 8 (308) (103), and if the input keyword is a special tag, the special tag (30) to rewrite the topic management table 7 with
7) Instruct the information management unit (103). The newly instructed dialogue flow knowledge is given a higher execution priority than the dialogue flow knowledge that has been executed until then, and execution is started.

If the keyword entered by the user is a special tag, the execution priority of the dialogue flow in the topic management table is rewritten accordingly (307) to explicitly execute the dialogue flow knowledge by the user. Even if no instruction is given, execution of the dialogue flow knowledge can be automatically switched.

For example, suppose that the user inputs the keyword “turn off the electricity” from the input / output means 9 while executing the dialog flow knowledge “turn on the electricity”. The keyword "turn off electricity" is transmitted from the input / output means 9 to the dialogue management unit 1 (901). The dialogue management unit 1 checks the keyword “turn off electricity” using the tag DB5 and the special tag DB6 managed by the knowledge management unit 2 (50).
2, 602). Since it is understood that the keyword "turn off electricity" is a special tag and represents a dialog flow name, a new parameter management table 8 is prepared, and the topic management table 7 includes a dialog flow "stop electricity". register. As a result, topic management table 7 and parameter management table 8
Are changed as shown in FIGS. 7 and 8, respectively.

When the dialogue management unit 1 finishes the interpretation process of the keyword input by the user, it continues to execute the dialogue flow knowledge. At this time, the topic management table 7 is changed as shown in FIG. Since the dialogue flow knowledge with the highest execution priority is "stop electricity", execution continues from the "start" node of the dialogue flow knowledge "stop electricity".

Also, "Turn off electricity" in the topic management table
Is written in the table number field of the dialog flow name, so the parameter management table of table number 2 is used to execute the dialog flow knowledge. However, the table number field of the topic management table can be rewritten according to an instruction from the user. For example, if the table number field of the dialog flow name “Turn off electricity” is rewritten to 1, then the “Turn on electricity” dialog flow knowledge that has been entered in the parameter management table of Table number 1 until then is executed. The obtained data can be reused.

As described above, the execution of the dialog flow knowledge can be switched.

When the dialogue management unit 1 finishes the execution of one dialogue flow, it instructs the information management unit 3 (103) to rewrite the execution priority of the dialogue flow in the topic management table 7 to the end state (307), Next, the dialogue flow name and node name of the dialogue flow having the highest execution priority are read (703, 3).
01), instructs the knowledge management unit 2 (102), reads the dialog flow knowledge (402, 201), and continues execution from the node name read from the topic management table 7. FIG. 9 shows the topic management table 7 when the execution of the "stop electricity" dialog flow is completed.

Next, consider a case where a call for another dialog flow knowledge is described in the dialog flow knowledge. The called conversation flow name is added to the topic management table 7, and the called conversation flow name is written in the parent topic field (103, 307). Depending on the user's input, the order of the dialog flow knowledge to be executed may be switched, and the execution of the called dialog flow may be performed before the called dialog flow knowledge. Even in such a case, it is possible to check whether the called conversation flow name is recorded in the topic management table, and if it is, whether the calling conversation flow name is written in its parent topic field. The execution order can be corrected so that the called party's interaction flow knowledge is executed first.

When the dialog management unit 1 tries to continue the execution of the dialog flow knowledge having the highest execution priority, a call of another dialog flow knowledge is described in a node to be executed next to the dialog flow knowledge. Suppose that For example,
Assume that the dialog flow knowledge of "turn on electricity" is executed, and then the process proceeds to the call of the dialog flow "power failure". The topic management table 7 is checked to see whether the called-up dialog flow name ("power failure") has already been recorded, and if so, its execution priority is checked (103, 307, 703, 301). ). If the execution priority has ended, the dialog management unit 1 considers that the execution of the node that calls another dialog flow knowledge has ended, and
Execute the next node. If it has not ended, the execution priority of the called interaction flow knowledge is set higher than the current interaction flow knowledge (103, 30).
7) The execution of the called-up dialog flow knowledge ("power failure") is prioritized over the calling-side dialog flow knowledge ("turn on electricity"). If the called conversation flow name ("power failure") is not recorded in the topic management table 7, the called conversation flow name ("power failure") is newly recorded (103, 307), and the execution is performed. Start.

When another dialog flow knowledge is called from within the dialog flow knowledge, the execution order of the dialog flow knowledge can be changed in accordance with the user's input. A case may occur where the execution priority is higher than the called conversation flow knowledge. For example, assume that the topic management table is as shown in FIG. The dialog management unit tries to execute the dialog flow knowledge of "turn on electricity" with a high execution priority, but the next node to be executed is the call of the dialog flow knowledge of "power failure". Check whether the dialog flow name "power failure" is written. "Blackout"
Since the dialog flow name has already been written, the execution priority is checked next, but the execution priority is "end"
By setting the execution priority of the dialog flow knowledge of "power failure" higher than the dialog flow knowledge of "turn on electricity", it is called "power failure" rather than the dialog flow knowledge of "turn on electricity". Prioritize execution of dialog flow knowledge.

When the data obtained during execution is corrected (901, 103, 308), the information management unit 3 checks the parent data field of the corrected data in the parameter management table 8 (308, 803). . Since data used as a search key for acquiring the data is recorded in the parent data field, data having the same parent data as the data to be corrected is checked (308, 80).
3) Check the parent data of the checked data (30)
8, 803) and checking data having the same parent data (308, 803).

Also, data having the corrected data as parent data is recursively checked. When the check is completed, the data with the check is deleted and the designated data is corrected (308, 803).

After that, the dialog flow knowledge using the parameter management table 8 in which the corrected data is recorded is
The search is performed by checking the table number field of the topic management table 7, and the node name field of the searched dialog flow name is rewritten to “start”. Although the dialog flow knowledge related to the corrected data will be re-executed, the data already recorded in the parameter management table 8 will not be re-executed and the user will not be asked the same question. Only necessary data can be obtained efficiently.

For example, a case where data is corrected while the parameter management table is as shown in FIG. 11 will be described. (90)
1) It is assumed that the value of the parameter name “customer number” has been corrected (103, 308).

The information management unit 3 checks the parent data field of “customer number” in the parameter management table 8 and obtains “Yamada” (308, 803). Since data used as a search key to acquire the data is recorded in the parent data field, data having the same parent data is checked (308, 803), and the parent data of the checked data is checked. (308, 803), and checks data having the same parent data (308, 803).
Repeat recursively.

Also, data having the corrected data as parent data is recursively checked. What is finally checked is data in which the parameter names are "name", "customer number", "address", and "telephone number". When the check is completed, the data with the check is deleted and the designated data is corrected (308, 803). After that, the dialog flow knowledge using the modified parameter management table
The search is performed by checking the table number field of the topic management table, and the node name field of the searched interaction flow is rewritten to “start”. In this way, only necessary data can be efficiently acquired and corrected.

[0054]

As described above, the user does not need to give an accurate instruction by inputting an approximate work purpose from several input / output means using several keywords and following the guidance based on the knowledge of the dialog flow. Can also receive work support.

Also, by rewriting the execution priority of the dialog flow knowledge in the topic management table according to the keyword input by the user, even if the user does not explicitly instruct the execution of the dialog flow knowledge, the automatic execution can be performed automatically. The execution of the dialog flow knowledge can be switched. The caller's conversation flow name is recorded in the topic management table, and if so, its caller's conversation flow name is written in its parent topic field to determine the callee's conversation flow name. The execution order can be corrected so that the execution of the dialog flow knowledge is performed first.

When the acquired data is modified, the related dialog flow knowledge is re-executed. For data already recorded in the parameter management table,
Since the application program is not re-executed or the user is not asked the same question repeatedly, only data that needs to be corrected can be efficiently acquired.

[Brief description of drawings]

FIG. 1 is a functional block diagram showing an embodiment of an interactive operation support system using the present invention.

FIG. 2 is a diagram showing a specific example of dialog flow knowledge according to the present invention.

FIG. 3 is a diagram showing a specific example of a tag DB according to the present invention.

FIG. 4 is a diagram showing a specific example of a special tag DB of the present invention.

FIG. 5 is a diagram showing a specific example of a topic management table according to the present invention.

FIG. 6 is a diagram showing a specific example of a parameter management table of the present invention.

FIG. 7 is a diagram showing another specific example of the topic management table of the present invention different from FIG. 5;

FIG. 8 is a diagram showing another specific example of the parameter management table of the present invention different from FIG.

9 is a diagram showing another specific example of the topic management table of the present invention, which is different from FIGS. 5 and 7. FIG.

FIGS. 5, 7, and 9 of the topic management table of the present invention.
It is a figure which shows another specific example different from FIG.

FIG. 11 is a diagram showing another specific example of the parameter management table of the present invention, which is different from FIGS. 6 and 8;

[Explanation of symbols]

 1 Dialog Management Section 2 Knowledge Management Section 3 Information Management Section 4 Dialog Flow Knowledge Base 5 Tag DB 6 Special Tag DB 7 Parameter Management Table 8 Topic Management Table 9 Input / Output Means 10 Knowledge Execution Section

Claims (4)

[Claims] An interactive system operation support apparatus for supporting a user by performing a task by giving appropriate guidance to a user who works by combining a plurality of application programs and commands of the application program. Input the work purpose as a keyword, provide guidance such as questions and explanations of operation methods, input / output means for presenting the obtained data to users, explanations of questions and operation methods to users, application programs Execution flow, a dialog flow knowledge base holding a plurality of dialog flow knowledge describing the condition judgment and the like using the obtained data, a tag DB that extracts parameter names and their values from the dialog flow knowledge base, A special tag that extracts tags used only in the specific dialog flow knowledge from the tag DB A knowledge DB that generates a tag DB from the dialog flow knowledge base, generates a special tag DB from the tag DB, and outputs the contents of the dialog flow knowledge base, the tag DB, and the special tag DB upon request. A knowledge execution unit that actually starts an application program or transmits a command sequence to the application program according to an instruction; and a dialog flow to be executed based on a result of a keyword input interpretation from a user or an execution result in the knowledge execution unit. A topic management table that records knowledge, an information management unit that manages the obtained data using a parameter management table that stores the set of attribute names and attribute values, and the tag DB via the knowledge management unit. Interpreting the keyword input by the user using the special tag DB,
The interpretation result is transmitted to the information management unit and recorded, and the knowledge is used by using the data recorded by the information management unit in accordance with the description of the plurality of interaction flow knowledge in the interaction flow knowledge base via the knowledge management unit. An interactive operation support device comprising: a dialog management unit that gives an instruction to an execution unit. 2. The tag DB is a tag DB in which a dialog flow name of the dialog flow knowledge, a parameter name used in the dialog flow knowledge and its value are extracted as tags, and the special tag DB is Special tag D which extracts tags used only for specific dialog flow knowledge from the tag DB
B, the keyword input by the user and the tag D
B. The interactive type according to claim 1, wherein a user's input interpretation, topic change, question generation to the user, and the like are performed by comparing the tag extracted with the special tag DB. Operation support device. 3. The topic management table, the execution number indicating the execution priority of the topic, the conversation flow name, the table number indicating the number of the parameter management table corresponding to the conversation flow, and up to what part of the conversation flow is executed. It is a topic management table having fields such as a node name that represents a parent topic and a parent topic that represents whether it is called from another interaction flow. The interactive operation support apparatus according to claim 1, wherein the execution order of the interactive flow is freely changed by recording the result. 4. The parameter management table according to claim 1, wherein the parameter management table has a parent data field for recording data used as a search key for acquiring the data obtained by the dialogue, in addition to the attribute name and the attribute value. 4. The interactive operation support device according to claim 1, wherein, by using the parameter management table, only data relevant when the data is corrected is extracted.