JP2611750B2 - Knowledge base search method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a knowledge base search method, and more particularly to a data group to be inferred (data group to be inferred).
If the limit value set in advance is exceeded during inference by performing inference on, the inference processing is interrupted and the interrupted alarm information is output, and the inference result so far is output to a file or the like. By doing so, unnecessary inference time can be deleted and alarm information can be analyzed, so how far the inference target data group was inferred,
The present invention relates to a knowledge base search method that facilitates determination of what is wrong with inference target data and performs effective inference processing.

[0002]

2. Description of the Related Art Next, a conventional knowledge base search system will be described in detail with reference to the drawings. FIG. 2 is a block diagram showing an example of a conventional knowledge base search method. FIG.
In the knowledge base search method shown in FIG. 1, an inference information input unit 11 for inputting an inference target data group and a limit value group, and an inference target data group and a limit value when an inference target data group is input from the inference information input unit 11. Inference control unit 1 by converting groups
3 and the preprocessing I / F unit 12 and the inference target data group and the limit value group are passed to the inference unit 14 to activate the inference unit 14 as many times as necessary for inference, and to pass the inference result group to the inference result output unit 16. The inference control unit 13, a knowledge base storage unit 15 storing various kinds of knowledge in advance, and inference based on the knowledge in the knowledge base storage unit 15 and the inference target data group received from the inference control unit 13. The inference unit 14 that passes the inference result group to the inference control unit 13
And an inference result output unit 16 for outputting to an output device such as an RT or a printer.

As an example of such a conventional knowledge base search method, there is a default value display setting method disclosed in Japanese Patent Laid-Open No. 3-263268.

[0004]

According to the conventional knowledge base search method described above, in the method in which the inference is continued even when the inference target data exceeds the limit condition, even if the inference target data does not excessively satisfy the condition. There is a disadvantage that useless processing and long-time search are performed because the inference processing is continued, and a preset default value is displayed by displaying a preset default value as in Japanese Unexamined Patent Publication No. 3-263268 default value display method. In the method that eliminates complicated processing and long-time search, only the default value is displayed and the search result up to the middle is discarded, so it is not possible to obtain information about where and what kind of failure occurred in the inference target data There is a disadvantage that.

[0005] An object of the present invention is to interrupt the inference processing when the limit value is exceeded, and to output the interrupted alarm contents, thereby eliminating the time for performing unnecessary inference processing.
It is an object of the present invention to provide a knowledge base search method capable of promptly confirming and correcting inference target data and, as a result, executing optimization of inference target data in a short time.

[0006]

According to a first aspect of the present invention, there is provided a knowledge base search method which defines data comprising a plurality of finite records to be inferred as an inference target data group and uses the data when forcibly terminating the inference processing halfway. A plurality of finite values to be defined is defined as a limit value group, various data indicating an inference state up to a certain point in time is defined as an inference state data group, and any of the inference state data groups defines the limit value group. Various alarm messages to be output in the case of exceeding are defined as an alarm information group, and a result obtained by executing an inference process based on the inference target data group and the limit value group is defined as an inference result group. Thus, (A) the inference information input unit for inputting the inference target data group, (B) the limit value input unit for inputting the limit value group, and (C) the inference target data group from the inference information input unit Once entered, Preprocessing for converting the inference target data group into a format suitable for inference and transmitting the same, and when the limit value group is input from the limit value input unit, converting the limit value group into a format suitable for inference and transmitting the same. An I / F unit, and (D) the inference target data group and the limit value group sent from the preprocessing I / F unit for inference, and the inference state data for a finite number of times required for inference. Inference is performed as long as any of the groups does not exceed the limit value group, and the inference result group is transmitted for outputting the inference result;
When the alarm information group is passed, an inference control unit that sends the alarm information group for outputting alarm information,
(E) a knowledge base storage unit that stores knowledge in advance, and (F) an inference based on the knowledge in the knowledge base storage unit and the inference target data group and the limit value group received from the inference control unit. Executing, passing the inference result group to the inference control unit, and, when any of the inference state data groups exceeds the value of the limit value group, the alarm information corresponding to the limit value is output to the inference result unit. (G) an alarm information output unit that outputs the alarm information group received from the inference control unit to an output device such as a CRT or a printer;
And (H) an inference result output unit that outputs the inference result group received from the inference control unit to an output device such as a CRT or a printer.

Further, the knowledge base search method according to the second invention defines a data comprising a plurality of finite records to be inferred as an inference target data group, and uses a plurality of data to be used when forcibly terminating the inference processing halfway. Define a finite number of values as a limit value group, define various data indicating the inference state up to a certain point in time as an inference state data group, and when any of the inference state data group exceeds the limit value group By defining various alarm messages to be output as an alarm information group, and defining a result obtained by executing an inference process based on the inference target data group and the limit value group as an inference result group,
(A) an inference information input unit that inputs the inference target data group and the limit value group, converts the data into a format suitable for inference processing, and transmits the converted data; and (B) the inference target input from the inference information input unit. The data group and the limit value group are sent for inference, and the inference is executed as long as any of the inference state data groups does not exceed the limit value group by a finite number of times required for inference, and the inference result group is An inference control unit that sends out the alarm information group for outputting the inference result and, when the alarm information group is passed, sends the alarm information group for outputting the alarm information; and (C) a knowledge base storing knowledge in advance. A storage unit, and (D) executing inference based on the knowledge in the knowledge base storage unit and the inference target data group and the limit value group received from the inference control unit, and inferring the inference result group. An inference unit that, when any one of the inference state data groups exceeds the value of the limit value group, sends alarm information corresponding to the limit value to the inference control unit together with the inference result group. (E) converting the alarm information group received from the inference control unit into a CR
An alarm information output unit for outputting to an output device such as a T or a printer; and (F) an inference result output unit for outputting the inference result group received from the inference control unit to an output device such as a CRT or a printer. It is composed of

[0008]

Next, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of a knowledge base search method according to the present invention. As shown in FIG. 1, the knowledge base search method according to the present embodiment uses the inference information input unit 1.
, Pre-processing I / F unit 2, inference control unit 3, and inference unit 4
A knowledge base storage unit 5, an inference result output unit 6, a limit value input unit 7, and an alarm information output unit 8.

The following describes the consistency of the product manufacturing process procedure.
An example is shown in which the present embodiment is used to automatically search and determine from a knowledge base in a batch format instead of in an interactive format with an operator.

First, the inference information input unit 1 shown in FIG. 1 performs inference target data (for example, a finite record having a finite field including a character string and a numerical value to be inferred).
Manufacturing process procedure data including multiple manufacturing process information)
Is entered.

The limit value input section 7 inputs a limit value group.

The preprocessing I / F unit 2 converts the inference target data input to the inference information input unit 1 into a format suitable for inference, and passes the data to the inference control unit 3.

The pre-processing I / F unit 2 includes a limit value group (for example, a limit value group including a plurality of limit values to be applied at the time of inference from the limit value input unit 7 and variable names of the limit values). In the manufacturing process procedure, if a process following a certain process is described, the data including the allowable number of errors when counting as one error and the variable name of the allowable number of errors is input. , Is passed to the inference control unit 3.

Therefore, the inference control unit 3 passes the inference target data group and the limit value group to the inference unit 4, and uses any one of the limit value groups by the necessary number of inferences in a batch format instead of an interactive format with the operator. The inference unit 4 is automatically activated as long as it does not exceed the inference result obtained by the execution of the inference unit 4 (for example, what kind of error has occurred in which manufacturing process and what reason The result of the determination of the consistency of the manufacturing process procedure including the result of whether an error has occurred or the absence of a defect, etc.) is passed to the inference result output unit 6 and alarm information (for example, which limit value is When the process number is exceeded, the inference control unit 3 passes the alarm information to the alarm information output unit 8.

The inference unit 4 includes a knowledge base storage unit 5
Inference (for example, inference of the consistency of the manufacturing process procedure) is performed based on the knowledge held in the inference and the inference target data group and the limit value group received from the inference control unit 3, and the inference result is transmitted to the inference control unit 3. When the information is passed and any one of the limit value groups is exceeded, the alarm information is passed to the inference control unit 3 together with the inference result.

The knowledge base storage unit 5 stores a plurality of finite knowledge bases in which knowledge to be inferred (eg, which manufacturing process is followed by which manufacturing process is regarded as an error) is described in advance. It is configured.

The inference result output section 6 outputs the inference result to an output device such as a CRT or a printer.

The alarm information output unit 8 is provided with the inference unit 4
The alarm information passed through the inference control unit 3 from the
Output to an output device such as T or a printer.

[0019]

As described above, according to the knowledge base search method of the present invention, when the limit value is exceeded, the inference processing is interrupted, and the interrupted alarm content is output.
The effect of eliminating the time for executing unnecessary inference processing and quickly confirming and correcting the inference target data can be achieved, and as a result, the inference target data can be optimized in a short time. have.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of a knowledge base search method according to the present invention.

FIG. 2 is a block diagram showing an example of a conventional knowledge base search method.

[Explanation of symbols]

 Reference Signs List 1,11 Inference information input unit 2,12 Preprocessing I / F unit 3,13 Inference control unit 4,14 Inference unit 5,15 Knowledge base storage unit 6,16 Inference result output unit 7 Limit value input unit 8 Alarm information output Department

Claims (2)

(57) [Claims] 1. A data consisting of a plurality of finite records to be inferred is defined as an inference target data group, and a plurality of finite values used when forcibly terminating the inference processing is defined as a limit value group. Various data indicating the inference state up to a certain point in time is defined as an inference state data group, and various alarm messages output when any of the inference state data group exceeds the limit value group are transmitted to the alarm information group. By defining the result obtained by executing the inference process based on the inference target data group and the limit value group as an inference result group, (A) inference information for inputting the inference target data group An input unit, (B) a limit value input unit for inputting the limit value group, and (C) a format suitable for inference when the inference target data group is input from the inference information input unit. Strange And a preprocessing I / F section for converting and transmitting the limit value group into a format suitable for inference when the limit value group is input from the limit value input section; The inference target data group and the limit value group sent from the processing I / F unit are sent out for inference, and any one of the inference state data groups exceeds the limit value group by a finite number of times required for inference. Execute inference unless otherwise, send the inference result group for output of inference result,
When the alarm information group is passed, an inference control unit that sends the alarm information group for outputting alarm information,
(E) a knowledge base storage unit that stores knowledge in advance, and (F) an inference based on the knowledge in the knowledge base storage unit and the inference target data group and the limit value group received from the inference control unit. Executing, passing the inference result group to the inference control unit, and, when any of the inference state data groups exceeds the value of the limit value group, the alarm information corresponding to the limit value is output to the inference result unit. (G) an alarm information output unit that outputs the alarm information group received from the inference control unit to an output device such as a CRT or a printer;
(H) an inference result output unit that outputs the inference result group received from the inference control unit to an output device such as a CRT or a printer. 2. A data consisting of a plurality of finite records to be inferred is defined as an inference target data group, and a plurality of finite values used when forcibly terminating the inference processing is defined as a limit value group. Various data indicating the inference state up to a certain point in time is defined as an inference state data group, and various alarm messages output when any of the inference state data group exceeds the limit value group are transmitted to the alarm information group. By defining a result obtained by executing an inference process based on the inference target data group and the limit value group as an inference result group, (A) the inference target data group and the limit value group And (B) sending the inference target data group and the limit value group sent from the inference information input unit for inference. Out of the inference state data group as long as any one of the inference state data groups does not exceed the limit value group, and sends the inference result group to output the inference result. When the group is handed,
An inference control unit for transmitting the alarm information group for outputting alarm information; (C) a knowledge base storage unit storing knowledge in advance; and (D) a knowledge in the knowledge base storage unit and the inference control unit. Performing inference based on the inference target data group and the limit value group received from
While passing the inference result group to the inference control unit, if any of the inference state data group exceeds the value of the limit value group, the alarm information corresponding to the limit value together with the inference result group, An inference unit to be passed to the inference control unit;
(E) an alarm information output unit that outputs the alarm information group received from the inference control unit to an output device such as a CRT or a printer; and (F) an inference result group received from the inference control unit. An inference result output unit for outputting to an output device such as a printer.