Matching system, matching method, and matching device

Abstract

The present invention aims to provide a technology that utilizes both structured and unstructured data. [Solution] The matching system of the present invention is a matching system that extracts record information similar to an arbitrary record information from a plurality of pre-recorded record information, with respect to record information including fixed data having a plurality of items and unstructured data which is in a free-text format. The matching system comprises a processor and a memory, the memory having: a fixed data search program that searches for record information containing fixed data similar to the fixed data of an arbitrary record information from a plurality of pre-recorded record information; an unstructured data structuring program that structures unstructured data and generates an unstructured data graph; and a structured data search program that searches for record information from the searched record information that has a graph structure similar to the graph structure of the unstructured data graph of an arbitrary record information.

Description

Technical Field

[0001] The present invention relates to a matching system, a matching method, and a matching device.

Background Art

[0002] In the healthcare field, an industrial physician may, for example, provide consultation (interview guidance, interview) for achieving work-life balance in conjunction with a health checkup. Although the examination items of a general health checkup are fixed, the viewpoints that need to be noted in the consultation are increasing year by year. At the same time, the types of support measures (measures in the workplace) that an industrial physician must propose are also increasing. Specific support measures include, for example, early detection of occupational diseases, countermeasures against overwork, mental health countermeasures, measures for ensuring the health of diverse workers, proactive health promotion measures, improvement of the business environment of the individual counselor, and the like. In order to propose a suitable one from various support measures to the counselor (support target), it is necessary to listen out the current work content of the counselor, workplace rules, etc., grasp the needs, and then make a judgment. However, it may be difficult for an inexperienced industrial physician to make a judgment. Such difficulty in consultation is not limited to the healthcare field, but is a common problem in general business of receiving consultations and proposing support measures, etc. Also, the proposal is not limited to presenting support measures, but also includes advice, feedback, and the like.

[0003] Therefore, there is a need for technology that supports counselors in formulating support plans when they receive consultations from clients. For example, Patent Document 1 proposes a technology that can evaluate implemented measures using basic resident data, long-term care insurance data, medical insurance data, and policy-specific data. Specifically, Patent Document 1 aims to "provide a community-based integrated care business system that enables accurate and quantitative analysis by appropriately setting various indicators using a unified database," and discloses the following as an invention of a community-based integrated care business system: "It has an indicator calculation means that calculates the values ​​of several pre-set indicators using a database in which basic resident data, long-term care insurance data, medical insurance data, and policy-specific data are unified. The indicator calculation means sets the user's O indicator as the duration until the stage of the user's physical and mental state changes. The user's P indicator is the results of the services received by the user during the duration. The S indicator is the results of the measures implemented by the service provider who provided services to the user during the duration." [Prior art documents] [Patent Documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2018-124960 [Overview of the project] [Problems that the invention aims to solve]

[0005] However, Patent Document 1 assumes that various types of data, namely basic data, long-term care insurance data, and medical insurance data, are managed in a standardized format. Furthermore, while Patent Document 1 shows how to propose support plans based on outputting the mode and standard deviation of responses to checklists and questionnaires, and the average duration of each service, it may be difficult to determine which support plan is appropriate for the individual client because all support measures relevant to the client's situation are displayed. To propose support measures that are tailored to the individual client's needs, it is important to utilize not only standardized data but also non-standardized data, such as descriptions of observations made by the counselor during the consultation. This invention has been made in view of the above points, and aims to provide a technology that utilizes both structured and unstructured data. [Means for solving the problem]

[0006] To achieve the above objective, a matching system according to one aspect of the present invention is a matching system that extracts recording information similar to an arbitrary recording from a plurality of pre-recorded recordings, with respect to recording information including fixed data having a plurality of items and unstructured data which is in a free-text format, wherein the matching system comprises a processor and a memory, the memory comprising: a fixed data search program that searches for recording information containing fixed data similar to the fixed data of an arbitrary recording from a plurality of pre-recorded recordings; an unstructured data structuring program that structures the unstructured data and generates an unstructured data graph; and a structured data search program that searches for recording information from the retrieved recordings that has a graph structure similar to the graph structure of the unstructured data graph of the arbitrary recording. [Effects of the Invention]

[0007] According to the present invention, both structured and unstructured data can be utilized. Other issues, configurations, and effects not mentioned above will be clarified by the description of the embodiments for carrying out the invention below. [Brief explanation of the drawing]

[0008] [Figure 1] Figure 1 is an illustrative diagram showing the process from when a counselor receives a consultation from a client who is a target of support, to when they propose support measures. [Figure 2] Figure 2 shows an example of the configuration of the matching system. [Figure 3] Figure 3 shows an example of the hardware configuration of a consultation support device. [Figure 4] Figure 4 shows an example of the functional configuration of the matching system. [Figure 5] Figure 5 is a sequence diagram showing an example of the processing performed in the matching system. [Figure 6] Figure 6 shows an example of an input format for consultation records. [Figure 7] Figure 7 shows an example of input data. [Figure 8] Figure 8 shows an example of the data stored in the consultation record database. [Figure 9] Figure 9 is a sequence diagram showing an example of the structuring process for the findings section. [Figure 10] Figure 10 shows an example of the structure of an influence dictionary. [Figure 11] Figure 11 shows an example of stored data in the findings column database. [Figure 12] Figure 12 is a sequence diagram showing an example of the findings column search process. [Figure 13] Figure 13 shows an example of a screen image displayed using the analysis results display program. [Figure 14] Figure 14 shows an example of a pop-up image displayed using the analysis results screen display program. [Figure 15] Figure 15 shows an example of a screen image displayed using the consultation record update program. [Figure 16]FIG. 16 is a diagram showing an example of a screen image updated using the consultation record update program.

BEST MODE FOR CARRYING OUT THE INVENTION

[0009] Hereinafter, embodiments will be described with reference to the drawings. In each drawing, the same or equivalent components and parts are given the same reference numerals. Also, the present invention is not limited by this embodiment.

[0010] In the present disclosure, the recorded information is information including regular data and irregular data. Here, the regular data is data organized according to a specific structure. Also, the irregular data is data not following a specific structure. In the present disclosure, as an example of the recorded information, the case of a consultation record as a record of a consultation response between an industrial physician and a consultee will be described. The consultation record may be filled in according to a specific format, and therefore, the data included in the consultation record can be distinguished into regular data and irregular data. Here, examples of the regular data include, for example, age, gender, evaluation by options, etc. Also, examples of the irregular data include data input by the user in a free description format in a findings column or the like. Note that, in the present disclosure, the case of a consultation record is described as the recorded information, but the present disclosure is not limited to this case and can also be applied to general information. Also, in the present disclosure, the case of data in the findings column is described as the irregular data, but the present disclosure is not limited to this case, and the present disclosure can also be applied to cases including data input in items other than the findings column. Also, in the present disclosure, the case of a consultation record between an industrial physician and a consultee is described, but the present disclosure is not limited to this case. The present disclosure is not limited to the healthcare field, and can be applied when giving advice, counsel, and proposals based on records of other communications and dialogues.

EXAMPLE

[0011] (Summary of the Invention) FIG. 1 is an illustrative diagram showing the flow from when a counselor receives a consultation from a client who is a support target until a support measure is proposed. In FIG. 1, a device interface 22, a consultation record DB 11, a findings column DB 12, and a consultation support apparatus 100 are shown, but details will be described later.

[0012] In FIG. 1, a scene where a client X who is a support target consults a counselor is shown. The counselor is, for example, an industrial physician. The counselor inputs the consultation record with the client X into the consultation support apparatus 100 via the device interface 22. The consultation record may be recorded based on a predetermined format, and the input data D1 input into the consultation support apparatus 100 includes fixed-form data indicating the age, gender, etc. of the client X, and non-fixed-form data which is the data of the findings column input in a free description format. In the present disclosure, based on the input data D1, the content of support for support targets similar to the client X is retrieved by referring to the consultation record DB (database) 11 and the findings column DB 12.

[0013] The consultation record DB 11 records a plurality of consultation records such as past consultation records conducted with the client X and consultation records conducted for other support targets other than the client X. The consultation support apparatus 100 performs a statistical similarity search with the client X from the consultation record DB 11 based on the input data D1. As a result of the statistical similarity search, for example, the support proposed for a support target with a high similarity to the client X and the number of support targets for which the support was proposed are shown.

[0014] [[ID=II]] In addition, in the findings column DB 12, the data of the findings column in the input data D1 is recorded in a structured (data structuring) state. Also, in the findings column DB 12, the data of the findings column in the past consultation records conducted with the client X and the consultation records conducted for other support targets other than the client X are also structured and then recorded. The consultation support apparatus 100 searches the findings column DB 12 for data similar to the data of the findings column of the input data D1, and extracts the consultation records of the support consultations corresponding to the retrieved data.

[0015] The consultation support device 100 searches the consultation record DB 11 and the observation column DB 12 and presents the counselor with consultation records similar to those of client X. In other words, the consultation support device 100 presents the counselor with consultation records of other support recipients for whom support measures recommended to be proposed to client X have been proposed, as well as graphs based on the data in the observation column of those consultation records (hereinafter also referred to as "observation column graphs"; details will be described later). By referring to the consultation records of support recipients other than client X, the client can understand, for example, that support B was proposed to one support recipient, or that support K was proposed to six support recipients. Furthermore, by referring to the observation column graphs, the client can understand whether the situation of client X is similar to that of other support recipients. By referring to the information output by this consultation support device 100, the counselor can propose support that is appropriate for client X.

[0016] Statistical similarity search in the consultation support device 100 can search for the most frequent support measure applied to many support recipients, but the support measures are only representative and do not search for support measures that are suitable for individual client X. On the other hand, the data recorded in the comments column reflects the situation and needs of client X at the time of consultation, and if the content of the comments column DB12 is similar, it is highly likely that the personal circumstances such as the situation and needs are also similar. By considering support measures in combination with interpersonal analysis (quantitative analysis), which is an analysis between support recipients performed using the consultation record DB11, and intrapersonal analysis (qualitative analysis), which is an analysis that reflects the situation of client X performed using the comments column DB12, counselors can propose support measures that are more suitable for client X. For example, in Figure 1, support B is proposed to one person and support K is proposed to six people. Statistically, support K is recommended, but if the comments column graph is also considered, support B may be judged to be a more appropriate proposal for client X's situation.

[0017] In Example 1, the user of the consultation support device 100 will be described as a counselor who directly interacts with the client at the consultation site. However, since the consultation support device 100 can be operated by anyone present at the consultation site, the client X may operate it directly, a third party other than the client X and the counselor may operate it, or two or more people, such as the client, the counselor, and a third party, may operate it simultaneously.

[0018] (System Configuration) Figure 2 shows an example of the configuration of the matching system 1000. The matching system 1000 is a system that extracts record information similar to an arbitrary record information from multiple pre-recorded record information, with respect to record information (consultation records) which include standardized data having multiple items and non-standardized data in the form of free-text. The matching system 1000 includes a consultation support device (matching device) 100 and an analysis device 200. The consultation support device 100 and the analysis device 200 are connected via a communication network 33. The consultation support device 100 includes a processor and memory, as will be described later, and is operated by, for example, a counselor who is receiving a consultation from a client. The analysis device 200 is capable of viewing multiple pre-recorded record information and is operated by a viewer (for example, another industrial physician, a lawyer, and another veteran counselor). The viewer views the consultation records entered by the counselor via the consultation support device 100 through the analysis device 200 and adds information such as support policies and advice. This disclosure describes the case where there is one consultation support device 100, but there may be two or more consultation support devices 100. Similarly, there may be two or more analysis devices 200. Furthermore, the matching system 1000 may include devices other than the consultation support device 100 and the analysis device 200.

[0019] (Hardware configuration) The consultation support device 100 comprises a processor 21 and memory (auxiliary storage device 25), and the memory contains a program that causes the processor 21 to execute various functions. Specifically, the consultation support device 100 can be implemented using a personal computer or a server. In Example 1, the consultation support device 100 is implemented using a personal computer 20 configuration.

[0020] Figure 3 is a block diagram showing an example of the hardware configuration of the consultation support device 100. The personal computer 20 comprises a processor 21, a device interface 22, a communication interface 23, a main memory 24, and an auxiliary memory 25, which are connected via a bus 26. The personal computer 20 has one of each component, but it may have multiple identical components.

[0021] The processor 21 is, for example, a CPU (Central Processing Unit) that executes processing according to a program stored in a storage device such as a hard disk. Alternatively, the processor 21 may be an electronic circuit (such as a GPU (Graphics Processing Unit)) that includes the computer's control and arithmetic functions. The processor 21 can perform arithmetic processing based on data and programs input from each component inside the personal computer 20 and output arithmetic results and control signals to each component. The processor 21 may also control each component of the personal computer 20 by executing the OS (Operating System) or applications of the personal computer 20.

[0022] The device interface 22 is an interface that connects to the input device 31 and the display 32 wirelessly or via a wired connection. The input device 31 includes a device for inputting various operation instructions to the personal computer 20, such as a keyboard, mouse, buttons, or touch panel. The display 32 includes a device for displaying a GUI (Graphical User Interface), such as a display or touch panel, to the user.

[0023] The communication interface 23 is an interface for connecting to the communication network 33 wirelessly or via a wired connection. The communication interface 23 can be one that conforms to existing communication standards. The personal computer 20 can exchange information with external devices 34 via the communication interface 23 and the communication network 33. External devices 34 include, for example, other consultation support devices and storage devices such as cloud storage.

[0024] The main memory 24 is a memory device that stores instructions executed by the processor 21 and various data, and the information stored in the main memory 24 is read by the processor 21. The main memory 24 is, for example, a volatile RAM (Random Access Memory) capable of storing information. Alternatively, the main memory 24 could be any electronic component capable of storing information.

[0025] The auxiliary storage device 25 is a storage device other than the main memory device 24, and like the main memory device 24, it can store instructions executed by the processor 21 and various data. The auxiliary storage device 25 is, for example, a non-volatile (non-temporary) hard disk capable of storing electronic information. Alternatively, the auxiliary storage device 25 may be any electronic component capable of storing information. For example, input data D1 entered by the counselor via the input device 31 is stored in the auxiliary storage device 25. Furthermore, as will be described later, the consultation content input program 1 can also be stored there.

[0026] Furthermore, the hardware configuration of the analysis device 200 is substantially the same as that of the consultation support device 100, and includes a personal computer, an input device, and a display.

[0027] Information from the consultation support device 100 is stored in the auxiliary storage device 25, and during execution, various data from the consultation support device 100 is written to the main storage device 24. Alternatively, information from the consultation support device 100 may be stored in, for example, the main storage device 24 or an external device 34. In other words, the consultation support device 100 is an information processing device having a CPU that executes processing according to a program stored in a storage device such as RAM or a hard disk.

[0028] Some or all of the programs and data that implement the processes described below may be stored in the auxiliary storage device 25 in advance, or, if necessary, may be stored in the auxiliary storage device 25 from other devices connected via the communication network 33 or from non-temporary storage media via the communication interface 23. Programs and data may also be stored in the auxiliary storage device 25 via an interface (not shown) provided by the personal computer 20. For example, the auxiliary storage device 25 stores the consultation content input program 1 and input data D1. Input data D1 includes the consultation record entered by the counselor. The programs and data stored in the auxiliary storage device 25 will be described later.

[0029] (Functional Configuration) Figure 4 shows an example of the functional configuration of the matching system 1000. The consultation support device 100 includes a consultation content input program (record information input program) 1, a standardized data search program 2, a findings field structuring program (non-standardized data structuring program) 3, a findings field search program (structured data search program) 4, an analysis result screen display program 5, a consultation record DB (record information DB) 11, a findings field DB (non-standardized data DB) 12, and an influence dictionary 13. These programs 1 to 5 and data 11 to 13 are stored in an auxiliary storage device 25 in advance, for example. The analysis device 200 also includes a consultation record search unit (record information search unit) 7, a consultation record update program (record information update program) 8, and an update result transmission unit 9.

[0030] The consultation content input program 1 is a program that allows the input of arbitrary record information (for example, a consultation record of the person being supported (the person seeking assistance)). When executed by the processor 21, the arbitrary record information is stored in the consultation record DB 11. The consultation content input program 1 includes control commands for the input device 31 and the display 32. The consultation content input program 1 displays the screen image during input on the display 32 using a GUI (Graphical User Interface), and also allows the counselor to input various operation instructions via the input device 31. For example, the consultation content input program 1 allows the counselor to input data to be processed via the input device 31, and also allows the counselor to input instructions to start processing on the consultation support device 100. If multiple consultation records are already stored in the consultation record DB 11, the consultation content input program 1 may allow the counselor to select the data to be processed from the data contained in the consultation record DB 11.

[0031] Standard data search program 2 is a program that searches for record information containing standard data similar to the standard data of a given record information from among multiple pre-recorded record information. By executing standard data search program 2, processor 21 searches for similar record information, which is record information containing standard data similar to the standard data of a given record information. Standard data search program 2 includes control instructions that cause processor 21 to search for standard data from the data contained in consultation record DB 11.

[0032] The unstructured data structuring program is a program that structures unstructured data and generates an unstructured data graph. When executed by the processor 21, it generates an arbitrary unstructured data graph, which is an unstructured data graph of arbitrary record information. The unstructured data DB stores the structured unstructured data. In this disclosure, to explain the case of data showing the industrial physician's findings when a consultation record is created, the unstructured data, unstructured data structuring program, and unstructured data DB are conveniently referred to as "findings column data," "findings column structuring program," and "findings column DB," respectively. The findings column structuring program 3 includes control instructions to the processor 21 to structure the data contained in the findings column DB 12. When data structuring is performed, the influence dictionary 13 is referenced in the findings column structuring program 3. The influence dictionary 13 is a database that stores terms that indicate predicates and influence values ​​associated with those terms. Details will be described later.

[0033] The structured data search program is a program that searches for similar graph record information, which is record information having a graph structure similar to the graph structure of the unstructured data graph of the arbitrary record information, from among the record information retrieved by executing the standardized data search program 2. The processor 21 searches for similar graph record information, which is record information having a graph structure similar to the graph structure of the arbitrary unstructured data graph, from among the similar record information by executing the structured data search program. For convenience, the structured data search program is also referred to as the "observation column search program". The observation column search program 4 includes control instructions that cause the processor 21 to search for specific information contained in the observation column DB 12.

[0034] The analysis results screen display program 5 is a program that displays similar recorded information, and when executed by the processor 21, a screen containing unstructured data associated with similar graph recorded information is displayed. The analysis results screen display program 5 includes control commands for the input device 31 and the display 32, and displays the results on the display 32 using a GUI or the like, allowing the counselor to perceive the processing results. In addition, the analysis results screen display program 5 allows the counselor to input various operation instructions via the input device 31.

[0035] Furthermore, the standard data search program 2 can be implemented on separate hardware, may be installed outside the consultation support device 100, or may be installed in a way that allows access via the communication network 33. Furthermore, the findings section structuring program 3 can also be implemented using separate hardware and may be installed outside the consultation support device 100. However, it should be installed in a way that prevents access via the communication network 33.

[0036] The analysis device 200 includes a consultation record search unit (record information search unit) 7 that searches for a predetermined consultation record (predetermined record information) from the consultation record DB 11 and a findings column graph (unstructured data graph) of a predetermined consultation record from the findings column DB (unstructured data DB) 4; a consultation record update program (record information update program) 8 that, when executed by the analysis device 200, adds information to the unstructured data of the predetermined consultation record and updates the predetermined consultation record; and an update result transmission unit 9 that transmits and stores the updated predetermined consultation record in the consultation record DB 11. Specifically, the consultation record search unit 7 can acquire data of the consultation record to be analyzed from the consultation support device 100 via the communication network 33. Furthermore, by the analysis device 200 executing the consultation record update program 8, viewers can add new findings and other data to the consultation record to be analyzed and update the data.

[0037] Furthermore, the analysis device 200 transmits data containing findings added by the viewer to the communication interface 23 via the communication network 33 using the update result transmission unit 9. The consultation record DB 11 and the findings column DB 12 store the updated consultation records received via the communication interface 23.

[0038] While the description has included a case where the matching system 1000 includes a consultation support device 100 and an analysis device 200, this disclosure is not limited to this case. The consultation support device 100 and the analysis device 200 may be implemented by hardware or by using computer resources such as the cloud. Furthermore, the consultation support device 100 and the analysis device 200 may be different devices or may be a single device.

[0039] (Matching system processing) Figure 5 is a sequence diagram showing an example of the process performed in the matching system. Steps S101 to 109 are performed in the consultation support device 100, and steps S110 to S112 are performed in the analysis device 200. The operation based on the flowchart in Figure 5 is as follows.

[0040] (Processing on the consultation support device 100 side) Step S101: The consultation support device 100 acquires the content of the consultation. The counselor conducts a consultation with the client and inputs the content of the consultation into the consultation support device 100. The input data D1 is stored in the consultation record DB 11 as a consultation record by the consultation content input program 1. The input data D1 includes standard data and non-standard data (remarks column data).

[0041] Step S102: The consultation support device 100 performs a search process for similar cases of non-standard data. The consultation support device 100 searches the consultation record DB 11 for consultation records (similar record information) that have standard data similar to the standard data contained in the input data D1 by executing the standard data search program 2.

[0042] Step S103: The consultation support device 100 performs a findings column structuring process. The consultation support device 100 structures the findings column data of the input data D1 and generates a findings column graph by executing the findings column structuring program 3.

[0043] Step S104: The consultation support device 100 performs a findings column search process. By executing the findings column search program 4, the consultation support device 100 searches for consultation records (similar graph record information) that have a graph structure similar to the graph structure of the findings column graph from among the similar consultation records found in step S102.

[0044] Step S105: The consultation support device 100 displays the search results. By executing the analysis results screen display program 5, the consultation support device 100 displays the consultation records found in step S104 on the display 32 as similar cases to the input consultation records.

[0045] Step S106: The consultation support device 100 determines whether or not to display the subsequent status of a similar case. The user selects whether or not to display the consultation record of the consultation that took place after the consultation record displayed in step S105. If the user chooses to display it (YES), the user proceeds to step S107; if the user chooses not to display it (NO), the user proceeds to step S108.

[0046] Step S107: The consultation support device 100 displays prognosis information on the screen, showing subsequent information for similar cases. The consultation support device 100 displays the prognosis information on the display 32 by executing the analysis results screen display program 5. After displaying the prognosis information, the process proceeds to step S105.

[0047] Step S108: The consultation support device 100 determines whether or not to suggest other similar cases. The user selects whether or not to display other similar cases regarding the consultation record displayed in step S105. The counselor refers to the similar cases displayed up to step S108 and proposes support measures to the client. The counselor enters the support measures and findings into the client's consultation record.

[0048] Step S109: The consultation support device 100 stores data in the consultation record DB 11 and the findings column DB 12. Suggestions for support measures made by the user are stored in the consultation record DB 11. The stored data is used as an example to search when other consultations are conducted. Suggestions for support measures may be stored as standardized data or as unstandardized data.

[0049] The consultation support device 100 can extract consultation records similar to the client's consultation record by performing statistical analysis in step S102 and qualitative analysis in step S104. Because it can refer to consultation records similar to the client's consultation record, the counselor can make suggestions tailored to the client's situation and needs while considering similar cases.

[0050] (Processing on the analysis device 200 side) Step S110: The analysis device 200 acquires consultation records. For example, when a viewer views a client's past cases, the consultation record search unit 7 of the analysis device 200 searches the consultation records. The analysis device 200 receives consultation record data from the consultation record DB 11 and the findings column DB 12 via the communication network 33. At this time, the viewer operating the analysis device 200 may be the counselor who proposed support measures to the support recipient in step S108, or it may be another counselor or another expert.

[0051] Step S111: The analysis device 200 displays updates to the consultation record. The consultation record acquired by the analysis device 200 is displayed, and the viewer performs processing to supplement non-standard data, such as adding observations to the consultation record. The analysis device 200 updates the consultation record processed by the viewer by executing the consultation record update program 8.

[0052] Step S112: The analysis device 200 transmits the update results to the consultation support device 100. The update result transmission unit 9 of the analysis device 200 transmits the updated consultation record to the consultation support device 100, and the transmitted consultation record is stored in the consultation record DB 11 and the findings column DB 12.

[0053] It is assumed that the viewer operating the analysis device 200 may be a third party different from the person seeking advice and the counselor. The counselor can refer to atypical data such as findings supplemented by the viewer via the consultation support device 100. Therefore, the counselor can make decisions on support measures after being aware of many different perspectives, making it possible to make suggestions that are more appropriate to the counselor's needs and circumstances.

[0054] (Details of the matching system's processing) The steps in the matching process described above will be explained with specific examples.

[0055] Step S101: Figure 6 shows an example of an input format for consultation records. Counselors input consultation records using the input format shown in Figure 6. The input format is displayed, for example, on the display 32 of the consultation support device 100. The data entered into the input format is classified into standardized data TD, which can be expressed as numerical data such as age, numerical data such as health checkup results, and categorical data, and non-standardized data AD, such as the findings column. In Figure 6, standardized data TD includes the name of the person being supported, the proposed support indicating the content of the support proposed to the person being supported, age, the results of the hearing test, and the results of the vision test. Note that standardized data TD is data that can only be entered in a predetermined format. For example, if a counselor mistakenly enters 25 years old in the support recipient column, the consultation support device 100 may notify the counselor that the input is incorrect. In addition, standardized data TD may be entered by the counselor using a pull-down menu.

[0056] Furthermore, non-standard data AD is data in the form of free-form text written in the comments section, such as "My major is noise research. I need to produce results by the end of this fiscal year, so I cannot take time off work." (Comments section data). Note that the classification of standard data TD and non-standard data AD is for convenience only and is not limited to the case of this disclosure. For example, proposal support is not limited to standard data, and consultants may be asked to input data in a free-form text format.

[0057] Figure 7 shows an example of input data D1. Input data D1 consists of data entered according to the consultation support format in Figure 6. Input data D1 includes the following items: "ID", "Support recipient", "Proposed support", "Age", "Hearing test [1000Hz]", "Hearing test [4000Hz]", "Visual acuity [left]", "Visual acuity [right]", "Date", and "Opinion". Of these, "ID" to "Date" are called the standard input data column 41, and data corresponding to standard data TD is entered. "Opinion" is called the non-standard input data column 42, and data corresponding to non-standard data AD is entered. The non-standard input data column 42 is a free-text opinion field data. "Support recipient" indicates the person who received consultation, and "Proposed support" indicates the type of support proposed to the support recipient. "Age" indicates the age of the person being described. "Hearing Test [1000Hz]" shows the results of a hearing test conducted at 1000Hz, and "Hearing Test [4000Hz]" shows the results of a hearing test conducted at 4000Hz. "Visual Acuity [Left]" shows the visual acuity of the left eye, and "Visual Acuity [Right]" shows the visual acuity of the right eye. "Date" indicates the month in which the consultation took place. "Opinion" shows the counselor's opinion regarding the consultation.

[0058] Furthermore, input data D1 may include data other than data entered using the input format shown in Figure 6. For example, it may include data extracted from basic data such as resident addresses held by the local government, data entered by the consultant X himself, or data converted into standardized data using named entity recognition or similar methods, such as data from statements made during the consultation.

[0059] Step S102: Figure 8 shows an example of stored data D2 that is stored in the consultation record DB11. The items in stored data D2 are the same as the items in input data D1 shown in Figure 7, and in addition to "Support target" and "Proposed support", it includes the input standard data column 43 ("Age", "Hearing test [1000Hz]", "Hearing test [4000Hz]", "Visual acuity [Left]", "Visual acuity [Right]", and "Date") and the input non-standard data column 44 ("Findings (opinion)"). Input standard data column 43 contains data that corresponds to standard data, and input non-standard data column 44 contains data that corresponds to non-standard data. The consultation record DB11 stores records of past consultation interactions with consultant X and records of consultation interactions performed with other support targets, and the data in each row of the consultation record DB11 corresponds to the input data for one consultation interaction.

[0060] In step S102, the consultation support device 100 executes the standard data search program 2 to search and extract consultation records from the consultation record DB 11 that contain input standard data columns 43 similar to input standard data columns 41 of the input data D1 entered in step S102. In the standard data search program 2, the determination of similar data is performed, for example, based on the cosine similarity of the standard data. Specifically, for the data in each row of input standard data columns 41 of the input data D1 and input standard data columns 43 in the consultation record DB 11, a vector W={w1,w2,…,w} is used. N The cosine value of the angle between two numerical vectors is calculated using the formula}. For example, the cosine value is calculated between the vector in the input standard data column 41 of input data D1 and the vector in the No. 1-1 portion of the input standard data column 43 of stored data D2, and between the vector in the input standard data column 41 of input data D1 and the vector in the No. 10-1 portion of the input standard data column 43 of stored data D2. Based on this, it is determined that the larger the cosine value, the greater the similarity to input data D1. In step S102, the extracted consultation records may, for example, be a predetermined number of consultation records selected from those with high cosine values, or consultation records with cosine values ​​equal to or greater than a predetermined value may be extracted.

[0061] When using vector representation, in order to process both numerical and categorical data simultaneously and generate a single unified vector, categorical data obtained in choice format is converted to numerical data using one-hot encoding, and these different formats of data are then integrated into a single vector.

[0062] In step S102, the similarity of consultation records is determined based on basic attributes of the consultant, such as age and health checkup results, included in the standardized data. Therefore, it is possible to search for records of consultations conducted in groups with similar attributes, such as the same age and similar health status, and to statistically analyze the proposed support in this group. In step S102, since it is sufficient to search for similarities among multiple standardized data for the input standardized data column 41 of input data D1, the search method is not limited to vectorization, and other methods may be used.

[0063] Step S103: (Details of the structuring process for the findings section (step S103)) Figure 9 is a sequence diagram showing an example of the findings section structuring process. Figure 9(a) shows the processing steps of the findings section structuring process, and Figure 9(b) shows a specific example of the processing in the processing steps. The consultation support device 100 executes the findings section structuring program 3 to extract subjects, predicates, and objects from the findings section data, searches for the influence values ​​of the extracted predicates from the influence dictionary 13, uses the extracted subjects and objects as nodes, uses the extracted predicates as edges connecting the nodes, and associates the searched influence values ​​with the edges to generate a findings section graph.

[0064] First, the consultation support device 100 obtains the input unstructured data column 42 from the input data D1 and the observation graph of the person to be supported from the observation column DB 12 (step S201). Specifically, the consultation support device 100 obtains the observation column data described in the input unstructured data column 42 of the input data D1. In addition, the consultation support device 100 obtains the past observation column graph of the person to be supported in the input data D1 from the observation column DB 12. At this time, if an observation column graph has not been created in the past consultation records of the person to be supported in the observation column DB 12, the observation column graph will not be obtained. In Example 1, the observation column data is analyzed and structured using natural language processing technology. The creation of the observation column graph in the observation column DB 12 will be described later.

[0065] An example of data in the comments section is shown in Figure 9(b), which reads, "My major is noise research. I need to produce results by the end of this fiscal year, so I cannot take time off work."

[0066] Next, the consultation support device 100 extracts triples from the findings data (step S202). Specifically, the consultation support device 100 clarifies the grammatical structure of each sentence in the findings data by applying morphological analysis and dependency structure analysis. Based on this analysis, it extracts triples from each sentence, each consisting of three elements: "subject," "predicate," and "object." For example, as shown in Figure 9(b), a triple containing "client" as the subject, "noise research" as the predicate, and "major" as the object is extracted. A triple containing "client" as the subject, "cannot rest" as the predicate, and "work" as the object is also extracted.

[0067] Next, the consultation support device 100 performs an influence determination for all triples extracted in step S202, determining whether or not the description of the "predicate" is listed in the influence dictionary 13 (step S203).

[0068] Figure 10 shows an example of the structure of the influence dictionary 13. The influence dictionary 13 includes "No.", "Predicate", and "Influence Value" as items. In the influence dictionary 13, each predicate is assigned an influence value that indicates either a positive or negative influence. For example, the predicate "to be completely cured" is assigned a positive influence of +1. Similarly, the predicate "to improve" is also assigned a positive influence of +1. On the other hand, the predicate "unable to rest" is assigned a negative influence of -1, and the predicate "to dismiss" is also assigned a negative influence of -1.

[0069] To explain the specific method of determining the impact, as shown in Figure 9(b), the consultation support device 100 converts the predicate from "improved" to its base form, such as "to improve". Then, it searches for "to improve" in the terms of the impact dictionary 13 and obtains the impact value of "to improve," which is "positive impact = +1". Similarly, if the predicate is "unable to rest," it searches for "unable to rest" in the terms of the impact dictionary 13 and obtains the impact value of "unable to rest," which is "negative impact = -1". If the predicate is not listed in the dictionary, it is determined to be neutral = 0.

[0070] Next, the consultation support device 100 structures the data of the influence value and triple (step S204). Specifically, the consultation support device 100 uses the "subject" and "object" of the triple extracted in step S202 as nodes, and connects the two nodes with an edge that indicates a relationship called a "predicate" to create a findings column graph. The consultation support device 100 can also associate the influence value obtained in step S203 with the Date value indicating the recorded month with the edge.

[0071] In the findings graph shown in Figure 9(b), the predicate is "Client" because it is the client X who is being supported, and the object is "Major." Also, since the predicate is "Noise Research," a graph is created connecting the nodes of "Client" and "Major" with an edge to "Noise Research." The edge is associated with a neutral influence value and a Date value of October. In addition, since the object for Client X may be "Work" and the predicate may be "Unable to take a break," a graph is created connecting the nodes of "Client" and "Work" with an edge to "Unable to take a break." In this graph, the negative influence value and a Date value of October are associated.

[0072] Furthermore, if a findings graph has been created in the past for the same client, the consultation support device 100 will add the newly structured content to the past findings graph, and if no past findings graph exists, it will create a new findings graph. However, the method of creating the findings graph is not limited to the above case, and other methods may be used. For example, any value from input data D1 to input standard data column 41 may be used as a node, and edges may be connected between the supporter and the node value, with the relationship being "corresponding column name of input standard data column 41".

[0073] Furthermore, in the observation structuring process, it is sufficient to extract nodes from the free-text data and obtain the relationships between nodes to create a graph (structure it). Therefore, the method for creating the graph is not limited to the method using the influence dictionary 13, and other methods for constructing relationships may also be applied.

[0074] Figure 11 shows an example of stored data D3 contained in the findings column DB12. The findings column DB12 includes the following items: "No.", "Node 1", "Node 2", "Relationship", and "Date". Each row of data in stored data D3 corresponds to one findings column graph. In the case of Figure 11, the findings column DB12 contains five findings column graphs. For example, in the case of ID 1, the nodes of "Consultant A" and "Hobbies" are connected by an edge with "Music Appreciation". The influence value is neutral, and the Date value is August.

[0075] Step S104: (Details of the findings search process (step S104)) Figure 12 is a sequence diagram showing an example of the findings column search process. First, the consultation support device 100 obtains findings column graphs corresponding to the consultation records extracted in step S102 from the findings column DB 12 (step S301). Specifically, the consultation support device 100 executes the findings column search program 4 to obtain data from the findings column DB 12 in which the "Support Target" and "Date" of the consultation records extracted in step S102 match the "Node1" and "Date" of the stored data D3, respectively. For example, if in step S102 of Figure 5, a similar consultation record is extracted as No. 10-1 in Figure 8, where the "Support Target" is Client B and the "Date" is October, then data No. 5 to No. 7, where "Node1" is Client B and the "Date" is October, are obtained from the stored data D3 of the findings column DB in Figure 11.

[0076] Next, the consultation support device 100 processes each of the observation column graphs acquired in step S301 one by one until it is processed in the next step S303 (step S302).

[0077] Next, the consultation support device 100 calculates the graph edit distance between the observation column graph of the support subject acquired in S103 and the observation column graph acquired in S301 (step S303). The graph edit distance (GED) is used to measure the similarity between two graph structures. GED is the minimum number of editing operations required to match the two observation column graphs, and there are three specific editing operations: "inserting a node or edge," "deleting a node or edge," and "changing the label of a node or edge."

[0078] The consultation support device 100 performs the process of calculating the GED for all the findings column graphs acquired in step S301 (step S304).

[0079] The consultation support device 100 acquires M (where M is a positive integer) consultation records that have a graph structure with a high similarity, determined based on the GED calculated in step S303. In this disclosure, M consultation records are acquired in ascending order of GED (step S305). By searching the consultation records based on the similarity of the graph structure, it becomes possible to qualitatively evaluate the similarity of free-text data.

[0080] (User interface in a matching system) Steps S105 to S109 in Figure 5 relate to the user interface between the consultation support device 100 and the counselor. Steps S110 to S112 in Figure 5 relate to the user interface between the analysis device 200 and the viewer. The following explanation will refer to the screen images.

[0081] Steps S105 and S106: Figure 13 shows an example of a screen image 81 displayed using the analysis results screen display program 5. The findings column display 810 shows the contents of the findings column for the case of client A (No. 1-1 in Figure 8). The findings column graph display 812 shows the findings column graph retrieved in step S104 of Figure 5. Here, the findings column graph including data from No. 5 to No. 7 in Figure 11 ("Node1" is client B, "Date" is October) is shown. In addition, the findings column display 813 shows the findings column for the case of No. 10-1 in Figure 8 ("Support Target" is client B, "Date" is October).

[0082] The findings column display 810 shows the contents of the findings column obtained in step S101, which states "Hobby: Listening to music at high volume." On the other hand, the findings column display 813 shows the contents "Only the person can operate the noise-emitting machine, and since it would halt the entire company's operations, they cannot take time off work." In addition, in the findings column graph display 812, predicates connecting nodes are written near the edges. The indication of influence value and Date value is omitted.

[0083] In step S105, the consultation support device 100 displays screen image 81 as the result of searching the consultation record. When the counselor is providing consultation to client A, they can refer to the findings column graph display 812 and findings column display 813 as similar cases.

[0084] If the counselor wishes to check the changes in the consultation record shown in the findings graph display 812 and findings display 813, they select the "Prognosis" button 84 (YES in step S106). If the "Prognosis" button 84 is not selected, the process proceeds to step S108.

[0085] Step S107: Figure 14 shows an example of a pop-up image 85 displayed using the analysis results screen display program 5. Pop-up image 85 is a screen that is overlaid on the screen image 81 shown in Figure 13. While the screen image 81 shown in Figure 13 showed a findings column graph with a Date value of October, the findings column graph display 850 in the pop-up image 85 in Figure 14 has "examination sound" added as a node, and an edge indicating "difficult to hear" is shown. In addition, the edge is labeled with "January" as the Data value and "adverse effect" as the Impact value, and the edge itself is highlighted compared to other edges. Findings column display 851 shows the content "Examination sound has become even more difficult to hear".

[0086] In step S107, the consultation support device 100 displays the pop-up image 85 in Figure 14 as prognosis information. The prognosis information used is data acquired after the data shown in the findings column graph display 812 and findings column display 813 in Figure 13. In this disclosure, the findings column graph display 850 uses the data for No. 8 in Figure 11 ("Node1" is client B, "Date" is January), and the findings column display shows the case of No. 10-2 in Figure 8 ("Support target" is client B, "Date" is January).

[0087] By displaying prognosis information, counselors can understand the changes in the recipient of support in similar cases and predict the impact on the person seeking assistance. When the user determines that they have fully grasped the prognosis information, they press the "close" button 86 to dismiss the pop-up image 85 and return to the screen image 81 in Figure 13. The consultation support device 100 then proceeds to step S105 and performs the following processing.

[0088] Step S108: If the user selects the "Research" button 83 in the screen image 81 of Figure 13, the consultation support device 100 removes the consultation record output to screen image 81 from the consultation records extracted in step S102 and proceeds to step S104. In this case, the screen image 81 displayed in step S105 shows information from other consultation records. When the consultation record review is complete and the counselor wants to proceed with processing a proposal for the consultant, the counselor selects the "Propose" button 82 (No. in step S106). When the "Propose" button is pressed, the "Proposal Support" item is obtained from the data of the corresponding person's consultation record DB 11 and observations column DB 12 and stored in input data D1. In this disclosure, when the "Proposal" button 82 is pressed, the data stored in the "Proposal Support" item (Support B) from No. 10-1 in Figure 8 is obtained and stored in the "Proposal Support" item of the input data D1 of consultant X shown in Figure 7.

[0089] When a counselor decides on a support measure to propose to the client, they enter it in the "Proposed Support" section of the consultation record format shown in Figure 6, and also make the necessary entries in the "Observations" section.

[0090] Step S109: The consultation support device 100 stores the information entered into the consultation record format in the consultation record DB 11 and the findings column DB 12.

[0091] Step S110: Figure 15 shows an example of a screen image 91 displayed using the consultation record update program 8. The findings column graph display 910 shows the findings column graph stored in the findings column DB 12 for a given support recipient. In Figure 15, the findings column graph is shown when Node 1 is client X and the Date value is August. The findings column display 911 shows the findings column data when the support recipient is client X and the Date value is August. The additional comments column 94 contains data entered by the viewer. The viewer refers to the findings column graph display 910 and the findings column display 911 and then enters their findings in the additional comments column 94 in a free-text format.

[0092] Steps S111 and S112: Figure 16 shows an example of a screen image 91 updated using the consultation record update program 8. When the viewer selects the findings update button 92, the content that was written in the additional information field 94 in Figure 15 is reflected in the findings field graph display 910 and the findings field display 911. The update result transmission unit 9 also transmits and stores the updated consultation record (findings field graph and findings field data) in the consultation record DB 11 and findings field DB 12 via the communication interface 23 and the communication network 33.

[0093] (Effects / Actions) As explained above, this disclosure allows the use of both structured and unstructured data. The matching system 1000 can use structured and unstructured data to extract consultation records similar to those of the person being supported. According to this disclosure, when a counselor receives a consultation from a person being supported, the matching system 1000 can extract a group of people with similar attributes to the person being supported, and then extract consultation records from that group whose descriptions in the comments section are similar. Because counselors can refer to consultation records of people being supported that are statistically and qualitatively similar to the person seeking advice before providing advice, they can make suggestions that are tailored to the person seeking advice's situation and needs.

[0094] Although embodiments of the present invention have been described above, the present invention is not limited to the embodiments described above, and various modifications are possible without departing from the spirit of the present invention.

[0095] The following describes, but is not limited to, embodiments that may constitute the present invention. (Aspect 1) A matching system for extracting record information similar to a given record information from multiple pre-recorded record information, with respect to record information including standardized data having multiple items and unstandardized data in the form of free-text descriptions, The aforementioned matching system, Processor and Equipped with memory, The aforementioned memory is A standard data search program that searches for standard data similar to standard data in a given record from multiple pre-recorded records, An unstructured data structuring program that structures the aforementioned unstructured data and generates an unstructured data graph, A structured data search program that searches for record information from among the retrieved record information that has a graph structure similar to the graph structure of the unstructured data graph of the arbitrary record information, A matching system that has the following features. (Aspect 2) The matching system described in Embodiment 1, The aforementioned memory is A record information database that stores the aforementioned multiple records of record information, An unstructured data database for storing the aforementioned structured unstructured data, A recording information input program that allows input of the aforementioned arbitrary recording information, The system further includes an analysis results screen display program that displays similar recorded information, The aforementioned processor, By executing the aforementioned recording information input program, the arbitrary recording information is stored in the recording information database. By executing the aforementioned standard data search program, similar record information, which is record information containing standard data similar to the standard data of the arbitrary record information, By executing the aforementioned unstructured data structuring program, an arbitrary unstructured data graph, which is an unstructured data graph of the arbitrary recorded information, is generated. By executing the structured data search program, similar graph record information is searched from among the similar record information, which is record information having a graph structure similar to the graph structure of the arbitrary unstructured data graph. By executing the aforementioned analysis results screen display program, a screen containing unstructured data associated with the similar graph recording information is displayed. Matching system. (Aspect 3) A matching system according to Embodiment 1 or Embodiment 2, The aforementioned standard data search program searches for recorded information based on the cosine similarity of the standard data. Matching system. (Aspect 4) A matching system according to any one of Embodiments 1 to 3, The aforementioned memory is The system further includes an influence dictionary that stores terms that represent predicates and influence values ​​associated with those terms. The processor executes the unstructured data structuring program, From the aforementioned unstructured data, the subject, predicate, and object are extracted. Search the influence value of the extracted predicate from the influence dictionary, The extracted subject and object are used as nodes. The extracted predicates are used as edges connecting the nodes. The unstructured data graph is generated by associating the searched influence values ​​with the aforementioned edges. Matching system. (Aspect 5) A matching system according to any one of embodiments 1 to 4, The similarity of the graph structures of the aforementioned arbitrary, unstructured data graphs is measured using graph edit distance. Matching system. (Aspect 6) A matching system according to any one of Embodiments 1 to 5, A matching device including the processor and the memory, Includes an analysis device capable of viewing a plurality of pre-recorded records, The aforementioned analytical device is A record information retrieval unit that searches for predetermined record information from the record information database and for unstructured data graphs of the predetermined record information from the unstructured data database, A recording information update program, which is performed in the aforementioned analysis device, adds information to the non-standard data of the predetermined recording information and updates the predetermined recording information, Includes an update result transmission unit that transmits and stores the updated predetermined record information in the record information database, Matching system. (Aspect 7) A matching system according to any one of embodiments 1 to 6, The aforementioned arbitrary record information includes records of consultations between the industrial physician and the person seeking advice. The aforementioned standardized data of arbitrary recorded information includes the results of a health checkup, The non-standard data of the aforementioned arbitrary record information is data indicating the occupational physician's findings at the time the arbitrary record information was created. Matching system. (Pattern 8) A matching method in a matching system that extracts record information similar to an arbitrary record information from multiple pre-recorded record information, with respect to record information including standardized data having multiple items and unstandardized data in the form of free-text descriptions, The aforementioned matching system, Processor and Equipped with memory, The aforementioned matching method is, From multiple pre-recorded records, search for records containing standard data similar to the standard data of a given record. The aforementioned unstructured data is structured, and an unstructured data graph is generated. The processor is instructed to perform the step of searching for record information from the retrieved record information that has a graph structure similar to the graph structure of the unstructured data graph. Matching method. (Aspect 9) A matching device for extracting record information similar to an arbitrary record information from multiple pre-recorded record information, with respect to record information including standardized data having multiple items and unstandardized data in the form of free-text descriptions, The matching device is Processor and Equipped with memory, The aforementioned memory is A standard data search program that searches for standard data similar to standard data in a given record from multiple pre-recorded records, A negative data structuring program that structures the aforementioned unstructured data and generates an unstructured data graph, A structured data search program that searches for record information from among the retrieved record information that has a graph structure similar to the graph structure of the unstructured data graph of the arbitrary record information, A matching device having [Explanation of Symbols]

[0096] 1…Consultation content input program 2…Standard data search program 3…Program for structuring the comments section 4…Research program for the findings section 5…Program for displaying analysis results screen 7…Consultation Record Search Department 8…Consultation Record Update Program 9…Update result transmission section 11...Consultation Record Database 12…Observations Database 13…Influence Dictionary 20… Personal computers 21… Processor 22…Device Interface 23…Communication Interface 24…Main memory 25…Auxiliary storage device 26... Pass 31…Input device 32…Display 33…Communication Networks 34...External device 100... Consultation and support device 200…Analyzer 1000... Matching System

Claims

1. A matching system for extracting record information similar to a given record information from multiple pre-recorded record information, with respect to record information including standardized data having multiple items and unstandardized data in the form of free-text descriptions, The aforementioned matching system, Processor and Equipped with memory, The aforementioned memory is A standard data search program that searches for standard data similar to standard data in a given record from multiple pre-recorded records, An unstructured data structuring program that structures the aforementioned unstructured data and generates an unstructured data graph, A structured data search program that searches for record information from among the retrieved record information that has a graph structure similar to the graph structure of the unstructured data graph of the arbitrary record information, A matching system that has the following features.

2. The matching system according to claim 1, The aforementioned memory is A record information database that stores the aforementioned multiple records of record information, An unstructured data database for storing the aforementioned structured unstructured data, A recording information input program that allows input of the aforementioned arbitrary recording information, The system further includes an analysis results screen display program that displays similar recorded information, The aforementioned processor, By executing the aforementioned recording information input program, the arbitrary recording information is stored in the recording information database. By executing the aforementioned standard data search program, similar record information, which is record information containing standard data similar to the standard data of the arbitrary record information, By executing the aforementioned unstructured data structuring program, an arbitrary unstructured data graph, which is an unstructured data graph of the arbitrary recorded information, is generated. By executing the structured data search program, similar graph record information is searched from among the similar record information, which is record information having a graph structure similar to the graph structure of the arbitrary unstructured data graph. By executing the aforementioned analysis results screen display program, a screen containing unstructured data associated with the similar graph recording information is displayed. Matching system.

3. The matching system according to claim 2, The aforementioned standard data search program searches for recorded information based on the cosine similarity of the standard data. Matching system.

4. The matching system according to claim 2, The aforementioned memory is The system further includes an influence dictionary that stores terms that represent predicates and influence values ​​associated with those terms. The processor executes the unstructured data structuring program, From the aforementioned unstructured data, the subject, predicate, and object are extracted. Search the influence value of the extracted predicate from the influence dictionary, The extracted subject and object are used as nodes. The extracted predicates are used as edges connecting the nodes. The unstructured data graph is generated by associating the searched influence values ​​with the aforementioned edges. Matching system.

5. The matching system according to claim 2, The similarity of the graph structures of the aforementioned arbitrary, unstructured data graphs is measured using graph edit distance. Matching system.

6. The matching system according to claim 2, A matching device including the processor and the memory, Includes an analysis device capable of viewing a plurality of pre-recorded records, The aforementioned analytical device is A record information retrieval unit that searches for predetermined record information from the record information database and for unstructured data graphs of the predetermined record information from the unstructured data database, A recording information update program, which is performed in the aforementioned analysis device, adds information to the non-standard data of the predetermined recording information and updates the predetermined recording information, Includes an update result transmission unit that transmits and stores the updated predetermined record information in the record information database, Matching system.

7. The matching system according to claim 2, The aforementioned arbitrary record information includes records of consultations between the industrial physician and the person seeking advice. The aforementioned standardized data of arbitrary recorded information includes the results of a health checkup, The non-standard data of the aforementioned arbitrary record information is data indicating the occupational physician's findings at the time the arbitrary record information was created. Matching system.

8. A matching method in a matching system that extracts record information similar to an arbitrary record information from multiple pre-recorded record information, with respect to record information including standardized data having multiple items and unstandardized data in the form of free-text descriptions, The aforementioned matching system, Processor and Equipped with memory, The aforementioned matching method is From multiple pre-recorded records, search for records containing standard data similar to the standard data of a given record. The aforementioned unstructured data is structured, and an unstructured data graph is generated. The processor is instructed to perform the step of searching for record information from the retrieved record information that has a graph structure similar to the graph structure of the unstructured data graph. Matching method.

9. A matching device for extracting record information similar to an arbitrary record information from multiple pre-recorded record information, with respect to record information including standardized data having multiple items and unstandardized data in the form of free-text descriptions, The matching device is Processor and Equipped with memory, The aforementioned memory is A standard data search program that searches for standard data similar to standard data in a given record from multiple pre-recorded records, A negative data structuring program that structures the aforementioned unstructured data and generates an unstructured data graph, A structured data search program that searches for record information from among the retrieved record information that has a graph structure similar to the graph structure of the unstructured data graph of the arbitrary record information, A matching device having

Images