JP2855133B2 - Expert system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an expert system for accumulating expert knowledge in the form of a knowledge base and performing inference processing using the knowledge base.

[0002]

2. Description of the Related Art FIG. 3 is a block diagram of a conventional example of an expert system, and FIG. 4 is a diagram showing the contents of a knowledge base 11 in FIG.

The expert system 2A is connected to the input / output device 1 in order to obtain external data, and manages an interface unit 10, a knowledge base 11 storing expert knowledge, and a knowledge base 11. A knowledge base management unit 12, an inference processing unit 13 for performing inference using the knowledge base 11, and a temporary storage unit 14 for storing input data or information obtained by inference processing using the knowledge base 11. It is configured.

The one element knowledge 31 of the knowledge base 11 is shown in FIG.
As shown in FIG.
And a group of conditional limbs 33 that define the conclusion 32. The condition limb 33 is composed of a judgment logic 34, an inspection item name 35, and a judgment value 36. An example 31 of one element knowledge of the knowledge base 11 is as follows.
It must be greater than 0. The second condition limb is
Similarly, the GPT is greater than 70, and the third conditional limb is that the TG is 270 or more. When the first, second, and third condition limbs are satisfied, one element knowledge 31 is established, and the hepatic dysfunction of the conclusion 32 indicates that the expert system is the inference result.

The inference processing unit 13 stores the data of each item input from the input / output device 1 via the interface unit 10 in the temporary storage unit 14, and then sequentially extracts one element knowledge 31 from the knowledge base 11. When the one-element knowledge 31 is established, the conclusion 32 is written to the temporary storage unit 14. When there is no more applicable elemental knowledge in the knowledge base 11, the conclusion part stored in the temporary storage unit 14 is determined as the final conclusion of the expert system 2A.

As described above, the knowledge base of the expert system has a plurality of knowledges.

In particular, the range of normal values for test values in the medical field differs depending on the test equipment and region used for the same test item. That is, the judgment logic 34 and the inspection item name 35 of the conditional limb 33 described in the one-element knowledge 31
Is the same, it is necessary to change the judgment value 36 according to the range of the normal value of the measuring device (inspection equipment) for the data input to the expert system and the region where the data is used.

For this reason, in the conventional expert system, it is necessary to construct a plurality of expert systems having a knowledge base described for each inspection device or a judgment value corresponding to a normal value of the environment in which the expert system is used. Was.

[0009]

However, it is cumbersome to create an expert system for each inspection device having a different judgment value, and when adding or deleting conditions to the knowledge base, a condition element of one element knowledge is required. (Decision logic 34
And the inspection item name 35) are irrelevant to the inspection device, so that the condition limb is added / deleted to each knowledge base of the expert system created for each inspection device, and the judgment value considering the normal value for each inspection device is added. I needed to edit.
In particular, the knowledge base is complicated, and managing and editing a plurality of knowledge bases not only takes time for management and editing, but also causes a trouble.

An object of the present invention is to provide an expert system capable of executing an inference process having a different judgment value for each inspection device, region, and the like by simpler and more reliable management of a knowledge base.

[0011]

An expert system according to the present invention comprises a reference value table storing reference values for each test item for describing and managing a knowledge base, and a normal value for each test item. A plurality of stored normal value tables, a data conversion unit that converts data input to the expert system based on the normal value table and the reference value table corresponding to the data, An inference processing unit that performs inference processing on the condition using the data converted by the data conversion unit.

[0012]

[Action] The data input to the expert system is
Since the data is converted based on the normal value table and the reference value corresponding to the inspection device or region where the data is obtained, only one complex knowledge base needs to be managed in the expert system.

[0013]

Next, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram of an expert system according to one embodiment of the present invention, and FIG. 2 is a flowchart showing the operation of the present embodiment.

The expert system of the present embodiment is configured by adding a reference value table 15, a normal value table 16, a data conversion unit 17, and a control unit 18 to the expert system of the conventional example of FIG.

The reference value table 15 defines the knowledge base 11, and is configured, for example, as shown in Table 1. The normal value table 16 is provided for each test device that has obtained input data from the input / output device 1, and stores a normal value for each test item. The control unit 18 controls the normal value table 16 relating to the data input from the input / output device 1 via the interface unit 10.
And the normal value table is selected. The data conversion unit 17 converts the input data into the selected normal value table 16.
And the reference value table 15. The knowledge base management unit 12 extracts applicable one element knowledge from the knowledge base 11. The inference processing unit 13 stores the data converted by the data conversion unit 17 in the temporary storage unit 14, and stores the data stored in the temporary storage unit 14 for each condition of one-element knowledge extracted by the knowledge base management unit 12. Perform inference processing using
The result is stored in the temporary storage unit 14.

[0017]

[Table 1] Next, the operation of this embodiment will be described with reference to the flowchart of FIG.

First, the control unit 18 controls the interface unit 10
, The identification number of the normal value table 16 relating to the data input from the input / output device 1 is input, and the normal value table (Table 2) is selected (step 21).

[0019]

[Table 2] Next, the control unit 18 determines whether there is data to be subjected to knowledge processing from the input / output device 1 via the interface unit 10 (step 22). If there is data to be subjected to the knowledge processing, the control unit 18
The data to be subjected to knowledge processing is input from the input / output device 1 (step 23). For example, assume that the data in Table 3 is input.

[0020]

[Table 3] Next, the data conversion unit 17 receives the input data (Table 3).
Is converted based on the normal value table 16 (Table 2) and the reference value table 15 (Table 1) selected in step 21 (step 24). The converted input data (Table 4) is stored in the temporary storage unit 14 via the inference processing unit 13.

[0021]

[Table 4] The data conversion in the data conversion unit 17 is performed by various methods depending on the characteristics of each inspection item.

Converted input data value = {(input data value-L1) (H2-L2) / (H1-L1)} + L
2. Next, the knowledge base management unit 12 extracts applicable one element knowledge from the knowledge base 11 (step 25). If there is no applicable one-element knowledge, the process returns to step 22 and the process is repeated. When there is applicable one-element knowledge, the inference processing unit 13 performs inference processing for each condition of the one-element knowledge using the data stored in the temporary storage unit 14 and stores the result in the temporary storage unit 14. Then, the process returns to step 25 to repeat the process (step 26). If there is no data to be subjected to the knowledge processing in step 22, the control unit 18 retrieves the conclusion in the temporary storage unit 14 via the inference processing unit 13 and sends the result to the input / output device 1 via the interface unit 10. Output (Step 27).

Inference processing based on input data obtained by another inspection device is performed without changing the knowledge base 11 by specifying a normal value table prior to the start of inference.

[0024]

As described above, the present invention has a normal value table for each inspection device or region, and converts input data using the corresponding normal value table and reference value table for the inspection device or region. By performing inference processing using the input data obtained,
Only one complex knowledge base is managed in the expert system, and inference processing with a knowledge base having different judgment values for each inspection device or region can be executed by preparing only a normal value table. Thus, there is an effect that an inference process having a different judgment value can be executed by simpler and more reliable management of a knowledge base.

[Brief description of the drawings]

FIG. 1 is a block diagram of an expert system according to an embodiment of the present invention.

FIG. 2 is a flowchart showing the operation of the embodiment of FIG.

FIG. 3 is a block diagram of a conventional example of an expert system.

FIG. 4 is a diagram showing the contents of a knowledge base 11 in FIG. 3;

[Description of Signs] 1 Input / output device 2 Expert system 10 Interface unit 11 Knowledge base 12 Knowledge base management unit 13 Inference processing unit 14 Temporary storage unit 15 Reference value table 16 Normal value table 17 Data conversion unit 18 Control unit 21 to 21 27 Step 31 One-element knowledge 32 Conclusion 33 Conditional limb 34 Judgment logic 35 Inspection item name 36 Judgment value

Continuation of front page (72) Inventor Kazuya Takeuchi 1-6-6 Uchisaiwaicho, Chiyoda-ku, Tokyo Nippon Telegraph and Telephone Corporation (56) References JP-A-2-75027 (JP, A) Proceedings of the Annual Meeting (late 1991) National Convention ”P. 2-183 to 184 (58) Fields investigated (Int. Cl. ⁶ , DB name) G06F 9/44

Claims (1)

(57) [Claims] 1. An expert system for accumulating expert knowledge in the form of a knowledge base and performing inference processing using the knowledge base, wherein a reference value for each test item for describing and managing the knowledge base is A stored reference value table; a plurality of normal value tables storing normal values for each of the test items; data input to the expert system; and a normal value table corresponding to the data; An expert system, comprising: a data conversion unit that converts based on a value table; and an inference processing unit that performs an inference process using data converted by the data conversion unit for each condition of one-element knowledge. .