DE102022105049A1 - METHOD, DEVICE AND SYSTEM FOR RECOMMENDED A KIT OF REHABILITATION PROGRAMS FOR A WORKER INJURED IN A WORKPLACE ACCIDENT - Google Patents

Abstract

The invention relates to an apparatus for recommending a set of rehabilitation programs for a worker injured in a work accident, which can help a work advisor who is a rehabilitation expert in rehabilitation planning by recommending a set of rehabilitation programs that is expected to be for most appropriate for workers injured in an accident at work. The apparatus includes a memory configured to store case data tables containing data on workers injured in a work accident who received rehabilitation service according to a set of rehabilitation programs and successfully returned to work, a search unit which, upon inputting attribute information, relating to a worker injured in a work accident and is configured to search the memory using the inputted attribute information and search for similar workers who have suffered similar accidents as the worker injured in a work accident, a similarity calculator used to calculate a similarity between the worker injured in a work accident and each of the searched workers who suffered similar work accidents, a control unit configured to select an injured worker having the greatest calculated similarity and is configured to generate a recommendation summary in which the rehabilitation programs included in a set of rehabilitation programs and provided to the selected worker with similar attributes are displayed according to the stages of medical care, and an output unit configured to output the recommendation summary.

Description

CROSS REFERENCE TO RELATED APPLICATION

This application claims priority from Korean Patent Application No. 10-2021-0030098 filed with the Korean Intellectual Property Office on March 8, 2021, the disclosure of which is incorporated herein by reference in its entirety.

BACKGROUND

1st area

The present application relates to a method, apparatus and system for recommending a set of rehabilitation programs to a worker injured in an industrial accident. More specifically, the present invention relates to a system and method for recommending a set of rehabilitation programs for a worker injured in a work accident, which can assist a labor counselor who is a rehabilitation expert in rehabilitation planning by recommending a set of rehabilitation programs, expected to be most appropriate for the worker injured in an industrial accident.

2. State of the art

An accident at work is a bodily injury suffered by an employee in an accident that occurs for a work-related reason, e.g. B. from the work environment or an action in the context of a work process. Workers' Injury Insurance is a social insurance provided by the government to promptly and equitably compensate workers after work accidents, rehabilitation and return of workers who have suffered a work accident (hereinafter "worker injured in a work accident") , to promote into society and to carry out projects and the like for the prevention of industrial accidents and the improvement of workers' welfare.

The Workers' Welfare and Compensation Service provides the Workers' Welfare and Compensation Service for workers' injuries, and the procedure for using the benefit is as follows. First, when a worker suffers an accident and applies to the Korea Workers' Welfare and Compensation Bureau for medical assistance, the Korea Workers' Welfare and Compensation Bureau will determine whether it is a labor accident, then whether or not the worker's compensation is granted medical assistance and communicated this decision to the employee. If the Korea Workers' Welfare and Compensation Bureau decides not to provide medical assistance, the employee can appeal the decision. If the Korea Workers' Welfare Bureau decides to provide medical care assistance, the worker can receive medical treatment at an industrial injury insurance medical facility (a medical facility designated by the Korea Workers' Welfare and Compensation Bureau or a public health center).

The benefits provided by the Korea Workers' Welfare and Compensation Bureau for workers who have suffered an industrial injury are generally divided into general benefits and case management (“find my job”) services. The general benefits are a set of benefits for workers injured in an industrial accident and are not part of the case management benefits. The case management services are services where a labor counselor (vocational rehabilitation coordinator or case manager) who is a rehabilitation expert provides nursing assistance service, medical, social psychological and vocational rehabilitation service, etc. on a one-to-one basis, so that a worker injured in an accident at work and having difficulties at his original place of work can prepare for his professional future.

The case management services include a process for counseling and assessment of the worker injured in an industrial accident, a process for determining the subject matter of the case management services based on the results of the counseling and assessment, a process for determining the ultimate goal of rehabilitation (return to original job, employment at another job and starting a business) and preparing a related rehabilitation plan, a procedure for providing a service according to the prepared rehabilitation plan, and a procedure for conducting follow-up management after the worker injured in a work accident returns to the original place of work.

The rehabilitation plan involves planning what rehabilitation program will be offered at what time etc. to the worker injured in an industrial accident. Depending on how appropriate the rehabilitation plans are for the worker injured in an industrial accident, the rate of return to work can vary.

Existing employment counselors rely mainly on their own experience when preparing rehabilitation plans. However, this is not optimal for inexperienced employment counselors, and even an experienced employment counselor takes a long time to create a rehabilitation plan.

SUMMARY OF THE INVENTION

The present invention aims to provide a method, an apparatus and a system for recommending a set of rehabilitation programs for a worker injured in a work accident, which can help a work counselor who is a rehabilitation expert in rehabilitation planning by recommending a rehabilitation program -Set recommends what is expected to be most appropriate for the worker injured in an industrial accident.

According to one aspect of the present invention, there is provided an apparatus for recommending a set of rehabilitation programs for a worker injured in a work accident, the apparatus including a memory configured to store case base tables including data on workers injured in a work accident have received a rehabilitation service according to a set of rehabilitation programs and have successfully returned to work, a search unit configured to search the store after inputting attribute information relating to a worker injured in an industrial accident and using the inputted attribute information and for workers seeks a similarity calculator used to calculate a similarity between the worker injured in a work accident and each that of the workers sought who have suffered similar industrial injuries, a control unit configured to select an injured worker with the greatest calculated similarity and to produce a recommendation overview in which rehabilitation programs included in a set of rehabilitation programs and with which the selected worker was provided with similar attributes corresponding to the stages of medical care, and an output unit configured to output the recommendation summary.

The case base tables can contain continuous attributes and nominal attributes.

The continuous attributes may include at least a severity index, age, average wage, number of days worked, number of workers injured at work, and a total score on a multidimensional psychological test.

The nominal attributes may include at least an injury and illness group code, an accident type code, an employee status code, an industry code, an occupation code, an indication of whether a workplace has been notified of a relevant accident, an indication of whether a relevant employee wishes to work at the same workplace, an an indication of whether the subject worker is married, a counseling target trend code, and an indication of whether a subject worker should be recommended a job.

The similarity calculator may include a first similarity calculator configured to calculate a Euclidean distance value according to each of the continuous attributes, a second similarity calculator configured to calculate a Hamming distance value according to each of the nominal attributes, and a summation unit configured to calculate a weighted Sum of Euclidean distance values and Hamming distance values is configured.

The summation unit may weight the Euclidean distance values using weight factors calculated separately on the continuous attributes and the Hamming distance values using weight factors calculated separately on the nominal attributes.

The inputted attribute information may include at least a worker injured worker severity index, an worker injured worker age, a worker injured worker average wage, a worker injured worker number of working days, a number of workers employed at a workplace injured worker injured worker total score on a multidimensional psychological test injured worker worker injured worker injury and disease group code injured worker accident type code employee status code at worker injured in a work accident, industry code of the worker injured in a work accident, occupation code of worker injured in a work accident, an indication of whether a workplace has the corresponding been notified of an accident, an indication of whether the worker injured in a work accident wishes to work at the same job, an indication of whether the worker injured in a work accident is married, a counseling target trend code, and an indication of whether the worker at a workers injured in an industrial accident should be recommended a job.

The search unit can search for case base tables that have the same injury and disease group code as the injury and disease group code.

The case base tables can be stored and managed in a mass data platform compiled by merging a labor insurance system of the Korea Workers' Welfare and Compensation Bureau with a medical information system of hospitals.

The mass data platform can have a star schema in which the case base table is considered as a fact table and tables selected in the employment insurance system are arranged around the case base table as dimension tables, a star schema in which two main tables selected in the medical information system are considered as separate fact tables, and others tables arranged as the two main tables around the main table as dimension tables, and a link table configured to link the case base table to the two main tables. The link table can contain a register number received from the case base table and a personal identification number received from the two main tables and an accident date as main attributes.

character list

• 1 ist ein Blockdiagramm eines Systems zur Empfehlung eines Sets von Rehabilitationsprogrammen gemäß einer beispielhaften Ausführungsform der vorliegenden Erfindung;
• 2 ist ein Diagramm zur Veranschaulichung eines Betriebsablaufs eines fallbasierten Begründungsmodels;
• 3A, 3B zeigen ein Sternschema bzw. Schneeflockenschema;
• 4 ist ein Diagramm zur Veranschaulichung eines Teils eines Logikmodells einer Massendatenplattform gemäß einer beispielhaften Ausführungsform der vorliegenden Erfindung;
• 5 ist ein Diagramm, das einige Tabellen zeigt, die von einem Arbeitsversicherungssystem aus den Tabellen ausgewählt werden, die eine Massendatenplattform bilden;
• 6 ist ein Diagramm, das einige Attribute einer Tabelle „Medizinische Erstversorgung“ unter den in 5 dargestellten Tabellen veranschaulicht;
• 7 ist ein Diagramm, das Tabellen zeigt, die von einem medizinischen Informationssystem aus Tabellen ausgewählt werden, die eine Massendatenplattform bilden;
• 8 ist ein Diagramm, das die Attribute einer Tabelle „Medizinische Informationen zur beruflichen Rehabilitation“ und einer Tabelle „Patienteninformationen zur beruflichen Rehabilitation“ veranschaulicht, die Haupttabellen unter den in 7 dargestellten Tabellen sind;
• 9 ist ein Diagramm, das eine Ausgabewerttabelle zum Speichern von Ausgabewerten eines Simulators veranschaulicht, der die Arbeitsfähigkeit eines bei einem Arbeitsunfall verletzten Arbeiters bewertet;
• 10 ist ein Blockdiagramm einer Vorrichtung zur Empfehlung eines Sets von Rehabilitationsprogrammen gemäß einer beispielhaften Ausführungsform der vorliegenden Offenbarung;
• 11 ist ein Diagramm, das 24 Rehabilitationsprogramme zeigt, die vom Koreanischen Amt für Arbeiterfürsorge und -entschädigung angeboten werden;
• 12 ist Diagramm, das 16 Attribute in einer Fallbasistabelle darstellt;
• 13 ist Entity-Relationship-Diagramm zwischen einer Rehabilitationsprogrammtabelle, einer Fallbasistabelle und Rehabilitationsinhaltstabellen;
• 14 ist ein Diagramm, das die Gleichungen veranschaulicht, die von einer Summierungseinheit verwendet werden, um eine endgültige Ähnlichkeit zu berechnen;
• 15A ist ein Diagramm, das ein Beispiel einer Empfehlungsübersicht zeigt, die durch eine Vorrichtung zur Empfehlung eines Sets von Rehabilitationsprogrammen angezeigt wird;
• 15B ist ein Diagramm, das ein weiteres Beispiel einer Empfehlungsübersicht zeigt, die durch eine Vorrichtung zur Empfehlung eines Set aus Rehabilitationsprogrammen angezeigt wird; und
• 16 ist ein Flussdiagramm, das ein Verfahren zur Empfehlung eines Sets von Rehabilitationsprogrammen für einen bei einem Arbeitsunfall verletzten Arbeiter gemäß einer beispielhaften Ausführungsform der vorliegenden Erfindung darstellt.

The above and other objects, features and advantages of the present invention will become more apparent to those skilled in the art through the detailed description of exemplary embodiments with reference to the accompanying drawings:

• 1 12 is a block diagram of a system for recommending a set of rehabilitation programs according to an exemplary embodiment of the present invention;
• 2 Fig. 12 is a diagram showing an operation flow of a case-based reasoning model;
• 3A , 3B show a star schema or snowflake schema;
• 4 Figure 12 is a diagram illustrating a portion of a logical model of a big data platform according to an exemplary embodiment of the present invention;
• 5 Fig. 13 is a diagram showing some tables selected by a labor insurance system from among the tables constituting a big data platform;
• 6 is a diagram showing some attributes of a Medical First Response table among the in 5 illustrated tables illustrated;
• 7 Fig. 12 is a diagram showing tables selected by a medical information system from tables constituting a mass data platform;
• 8th Figure 12 is a diagram illustrating the attributes of a Medical Rehabilitation Vocational Information table and a Patient Rehabilitation Vocational Information table, the main tables among those in 7 Tables shown are;
• 9 Fig. 12 is a diagram illustrating an output value table for storing output values of a simulator that evaluates work ability of a worker injured in a work accident;
• 10 12 is a block diagram of an apparatus for recommending a set of rehabilitation programs according to an exemplary embodiment of the present disclosure;
• 11 Fig. 12 is a chart showing 24 rehabilitation programs offered by the Korea Workers' Welfare and Compensation Bureau;
• 12 Fig. 12 is a diagram showing 16 attributes in a case base table;
• 13 is entity-relationship diagram between a rehabilitation program table, a case base table, and rehabilitation content tables;
• 14 Figure 12 is a diagram illustrating the equations used by a summation unit to calculate a final similarity;
• 15A Fig. 12 is a diagram showing an example of a recommendation map displayed by a rehabilitation program set recommendation apparatus;
• 15B Fig. 12 is a diagram showing another example of a recommendation map displayed by a rehabilitation program set recommendation apparatus; and
• 16 12 is a flow chart illustrating a method for recommending a set of rehabilitation programs for a worker injured in an industrial accident according to an exemplary embodiment of the present invention.

DETAILED DESCRIPTION

The advantages and features of the present invention and the methods of implementing it will become clearer with reference to exemplary embodiments described in detail with reference to the accompanying drawings. However, the present invention is not limited to the exemplary embodiments disclosed herein and can be implemented in various forms. The exemplary embodiments are provided only so that this disclosure will be thorough and complete, and will fully convey the scope of the present invention to those skilled in the art. The present invention is only defined by the scope of the claims.

Unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as is commonly understood among those skilled in the art. Also, terms defined in current dictionaries are not to be construed in an idealized or overly formal sense unless they are clearly so defined here.

The terms used herein are used only to describe the exemplary embodiments and not for purposes of limitation. The singular includes the plural, unless the context clearly indicates otherwise. As used herein, the terms "comprises" and/or "comprising" do not exclude the presence or addition of one or more elements other than those specified.

Exemplary embodiments of the present invention will be described below with reference to the accompanying drawings. In the drawings, like reference numbers refer to like elements.

1 12 is a block diagram of a system 1 for recommending a set of rehabilitation programs for a worker injured in a work accident (hereinafter “rehabilitation program recommendation system”) according to an exemplary embodiment of the present invention.

As 1 1, the rehabilitation program set recommendation system 1 comprises a mass data platform 100 and a rehabilitation program set recommendation device 200.

The big data platform 100 is constructed by merging a labor accident insurance and compensation system (hereinafter, “labor insurance system”) of the Korea Workers' Welfare and Compensation Bureau with an operating system of hospitals (hereinafter, “medical information system”). The data 110 extracted from the work insurance system and the data 120 extracted from the medical information system are stored and managed in the mass data platform 100 . The mass data platform 100 is described below with reference to the 3A until 9 described.

The rehabilitation program set recommendation device 200 operates on the basis of the big data platform 100. Specifically, the rehabilitation program set recommendation device 200 receives attribute information on a worker injured in a work accident from a job consultant, and using the received attribute information searches for a rehabilitation program set from which to expect that it is most suitable for the worker injured in a work accident. After that, the rehabilitation program set recommending device 200 configures and outputs a recommendation summary with a found rehabilitation program. Then, the labor counselor can make a rehabilitation plan for the worker injured in a work accident by referring to the content of the issued recommendation sheet. The job counselor can create a rehabilitation plan using the set of rehabilitation programs as displayed in the recommendation overview, or create a rehabilitation plan by modifying a rehabilitation program type and/or a delivery period.

According to the exemplary embodiment, the rehabilitation program set recommendation device 200 recommends a set of rehabilitation programs for the worker injured in a work accident based on a case-based reasoning model (CBR). The CBR model is an artificial intelligence (or machine learning) model that operates by mimicking human problem-solving methods.

CBR is a technique that looks for a similar case to a new case in the case base bank storing previous cases and predicts an outcome for the new case based on the similar cases retrieved. CBR is known to be very similar to human problem-solving methods. When a new problem to be solved arises, people think of the most similar case to the problem to be solved from among the problems they have solved in the past, modify the then-solution according to the current problem, and then use the modified solution as solution to the new problem. If the result indicates that the changed solution can be used in the future, people tend to remember the changed solution and use it again in a situation where a different problem needs to be solved. The CBR model works according to a similar process used in 2 is shown.

As in 2 shown, the CBR model essentially works in four phases. The first phase is a retrieval phase. In the retrieval phase, from a large number of cases stored in a case base bank, the one that most closely resembles a new problem that is currently to be solved is selected. In this case, only the most similar case can be retrieved, or a plurality of cases can be retrieved and matched. The second phase is a reuse phase. In the reuse phase, the case found in the previous phase is reused as the solution to the new problem. The third phase is the control phase. As a result of reusing the solution of a similar case from the past as the solution to the new problem, it may be necessary to change or adjust the solution slightly before saving the new case. In this case, the new case and the solution will be adjusted or changed in the control phase. Finally, the fourth phase is a storage phase. In the storage phase, the new resolution case, adjusted or changed as needed, is stored in the case base bank. Storing the new case in the case base bank along with the new solution means there is another way to solve a new problem. If such storage of new cases is repeated over a longer period of time, many new cases will accumulate and the performance of the CBR model will correspondingly become more accurate and precise. Therefore, this storage phase is considered the “learning” phase in the CBR. These four phases are referred to as the "4 REs"("REtrieve" - retrieve, "REuse" - reuse, "REvise" - control, "REtain" - store) of case-based reasoning.

Meanwhile, the data of the existing labor insurance system and the data of the medical information system have completely different forms and content. For example, the tables related to the employment insurance system use a "registration number" as a key, while the tables related to the medical information system use a "personal identification number" and an "accident date" instead of a "registration number" as a complex key. Accordingly, in order to search for data on a specific worker injured in a work accident, it is necessary to separately access the labor insurance system and the medical information system. On the other hand, if a mass data platform is used in which the tables of the two systems are linked, it is possible to easily and quickly search for data on the worker injured in a work accident.

According to the exemplary embodiment, a big data platform may be designed according to a data warehouse (DW) design technique. When a big data platform is designed according to the DW design technique, the labor insurance system and the medical information system with different forms and contents can be distinguished and appropriately integrated into one system.

Examples of the DW design technique can be modeling using a star schema and snowflake schemas.

Modeling using star schemas is the most basic DW design technique. A star schema includes a fact table and one or more dimension tables. The fact table is in the center and the one or more dimension tables are arranged in a star shape around the fact table. The fact table is configured to collect keys of the dimension tables. In other words, the star schema has a form in which the data can be easily summarized and viewed by dimensions of interest. 3A shows an example of a star schema.

When modeling using a snowflake schema, multiple star schemas are configured contiguously. A snowflake schema has a shape in which dimension tables are arranged around a fact table and other dimension tables are related to the dimension tables. 3B shows an example of a snowflake scheme.

According to the exemplary embodiment, the big data platform can be designed in a form similar to a snowflake scheme. This will be discussed in more detail with reference to 4 described.

4 FIG. 12 is a diagram showing a portion of a logical model of a big data platform according to an exemplary embodiment of the present invention. Tables related to the labor insurance system are in 4 shown on the left and tables related to the medical information system are shown on the right.

In the left part of 4 are the tables related to the employment insurance system (e.g. a table “Virtually Reality (VR) assessment”, a table “First Aid Medical Care”, a table “Time” and a table “Data Panel on Work Injury”) , arranged in a star schema around a Case Base table as a fact table. Also, a key attribute of the "Case Base" table is a register number.

In the right part of 4 are tables related to the medical information system (e.g., a daily activity test table, a balance assessment table, a shoulder multidimensional assessment table, and a hand function test 1 table), in the form of a star schema around two main tables ( e.g., a "Main Medical Information" table and a "Main Patient Information" table) arranged around them as separate fact tables.

The case base table, which is the fact table of the labor information system, is linked to the two main tables, which are the fact tables of the medical information system, by a link table, and thus all the tables are arranged in the form of a snowflake schema. The main attributes of the link table are also a personal identification number (encrypted), an accident date and a register number.

In addition to the in 4 In addition to the tables shown, the mass data platform includes many other tables. In particular, the big data platform contains 72 tables selected from the labor insurance system and 16 tables selected from the medical information system. In addition, the big data platform may also include an output table for storing output values of a simulator (not shown) that evaluates the skills of the worker injured in a work accident. The tables that make up the Big Data Platform and a main attribute of each table are described below with reference to the 5 until 9 described in more detail.

5 Fig. 12 is a diagram showing some tables selected by an employment insurance system from among the tables constituting a big data platform. In 5 45 tables of the 72 tables selected by the labor insurance scheme are presented. Below the presented tables are also tables with “Accident Information”, “Cause and Circumstances of Each Accident”, “Injury and Illness Code”, “Information on Payment of Medical Expenses”, “Pharmacy of Medical Institution”, “Data on Payment of Benefits Temporary Incapacity or Injury and Illness, Nursing Assistance Payment Details, Disability Payment Details, Insured Person Data Status, National Pension Acquisition Data, Health Insurance Normalization Table , "Detailed treatment report", "Details of the initial consultation card", "Rehabilitation counseling register", "Information on the calculation of vocational training costs", "Need for vocational training costs", "Details on the need for vocational training costs", "Details on the calculation of vocational training allowance", "Details on the need for vocational training allowance" and "Branch office" used to indicate a case basis to form a table. In particular, the case base table is configured to contain attributes selected from among the attributes contained in the 20 tables according to a specific reference.

6 is a diagram showing some attributes of a Medical First Aid table among the in 5 shown tables shows. The table for first aid medical care is prepared when a worker who suffers an industrial injury first informs the Korea Workers' Welfare and Compensation Bureau of the accident and applies for recognition of the industrial injury. Accordingly, the table for primary medical care contains most of the data from the application for recognition. Regarding 6 contains the table for medical first aid contains 123 attributes. In the table for medical first aid, a "registration number" serves as a key and data about the applicant, such as: B. a customer identification (ID), a registration number of the workplace, an accident date and a job title, as well as data on the workplace where the accident occurred are managed together. In 6 "PS" stands for "primary key" and "NN" for "non-zero" (an essential value).

7 Fig. 12 is a diagram showing the tables selected by a medical information system from among the tables constituting a mass data platform. 7 shows two main tables (medical information on vocational rehabilitation and patient information on vocational rehabilitation) and 14 tables. The 14 tables are mainly tables for managing various tests and test result values.

8th Figure 12 is a diagram illustrating the attributes of the Medical Rehabilitation Vocational Information table and the Patient Rehabilitation Vocational Information table, which are the main tables of the in 7 tables shown are. The table with medical information on vocational rehabilitation contains mainly information about medical facilities and medical personnel. For example, the Rehabilitation Medical Data table contains attributes such as Hospital Name, Medical Department, and Author (Physician). The occupational rehabilitation patient information table mainly contains information about the worker injured in a work accident and about accidents. For example, the occupational rehabilitation patient information table contains attributes such as “name” of the worker involved in the accident, “accident date”, “age” and “gender”.

9 Fig. 12 is a diagram showing an output value table for storing output values of a simulator that evaluates work abilities of a worker injured in a work accident. As in 9 shown, the output value table contains information based on evaluation results such as scores and angular lines. The output value table includes a total of 38 attributes. Two of the 38 attributes can be used as a complex key. The simulator that evaluates the work abilities of the worker injured in a work accident may be an augmented reality (XR) providing device including virtual reality (VR), augmented reality (AR), and mixed reality (MR).

10 12 is a block diagram of the rehabilitation program set recommendation device 200 according to an exemplary embodiment of the present invention.

As in 10 1, the rehabilitation program set recommending device 200 includes a memory 210, an input unit 220, an output unit 230, a control unit 240, a search unit 250, and a similarity calculator 260.

The memory 210 stores various pieces of data required for the operation of the rehabilitation program set recommender 200 . For example, the memory 210 stores a rehabilitation program table 213, a case base table 211, and a rehabilitation detail table 212 which links the rehabilitation program table 213 and the case base table 211 together.

The rehabilitation program table 213 stores data on rehabilitation programs provided by the Korean Workers' Welfare and Compensation Bureau. The Korean Workers' Welfare and Compensation Bureau offers a total of 24 rehabilitation programs related to the fields of medicine, social psychology and occupation, sociopsychology and occupation, with rehabilitation programs in 11 are shown.

As 11 shows, in the field of medicine, rehabilitation programs are offered for intensive rehabilitation treatments (special treatments), medical care support consultation, general practitioner consultation, rehabilitation sports, complication prevention and management, and rehabilitation aids, etc. In the field of social psychology, rehabilitation programs are offered for psychological counseling, family counseling, psychological rehabilitation programs, mentoring programs, family harmony programs, welfare applications, etc. With regard to the area of the profession, rehabilitation programs for support counseling on return to work, support support on employment in another job, assessment of work ability, vocational training, registration for job search (WorkNet), outsourcing of employment services, support support on the Business start-up, return-to-work grants, rehabilitation exercise costs, workplace adaptation training costs, projects to support substitute workers, programs to promote harmony among employees, etc.

At this point, on 10 which shows how the case base table 211 stores data about each case. According to an exemplary embodiment, the case base table 211 includes a total of 16 attributes that are 12 are shown.

12 Figure 12 shows the division of attributes of the case base table 211 into a continuous type and a nominal type. The continuous type includes 6 attributes and the nominal type includes 10 attributes. The continuous attributes (or continuous variables or numeric variables) include a severity index, age, average wage, number of days worked, number of workers injured at the same job, and a total score on a multidimensional psychological test. The nominal attributes (or nominal variables or categorical variables) include an injury and illness group code, an accident type code, an employee status code, an industry code, an occupation code, an indication of whether the workplace has been notified of a relevant accident, an indication of whether a relevant worker would like to work at the same place of work, an indication of whether the corresponding worker is married, a counseling target trend code and an indication of whether one appropriate worker should be recommended a job.

In the 12 Attributes presented are determined by a process of selecting candidate attributes based on two different methods, a process of analyzing the importance of the selected candidate attributes and a process of selecting a final attribute based on the analyzed importance.

Examples of the two methods of selecting candidate attributes can be a data-driven approach and an expert-driven approach. A process for selecting candidate attributes based on the data-driven approach includes a procedure for selecting 356 candidate attributes firstly from the attributes of 960 tables constituting the existing employment insurance system and a procedure for selecting 156 candidate attributes secondly from the first selected candidate attributes. A process for selecting candidate attributes based on the expert-driven approach, i. H. based on the Delphi technique, involves a method of organizing a panel that includes a variety of experts, a first round of conducting a survey in which open-ended questions are asked of the panel, a second round of conducting a survey that is based on based on a first round outcome and in which closed-ended questions are asked of the panel, a third round of conducting a survey by providing responses from other panels and structured closed-ended questions to the panel, a method of conducting a poll by providing responses other panels and structured closed-ended questions to the panel until the panel's opinions converge, and a process of determining 35 candidate attributes based on the poll results.

A light gradient amplifier machine (LGBM) is used to analyze the importance of 191 candidate attributes selected in this way. The LGBM is configured to take candidate attributes as the independent variable and a rehabilitation program as the dependent variable. The LGBM calculates the importance of the candidate attributes based on each candidate attribute's contribution to improving accuracy and node distribution. As a result, 16 attributes are finally selected from the 191 candidate attributes.

coming back on 10 the memory 210 stores case base tables 211 relating to various example cases. Example cases include at least five rehabilitation programs from sets of rehabilitation programs offered to any worker injured in an industrial accident who successfully returned to work. Return to work here includes returning to the original job, returning to work at another job and starting a business.

The example cases may include actual cases and virtual cases. Virtual cases can be generated by selecting an injury and a disease, an injured body part and an age group, which are rare in real cases, by providing the job advisor with information about a virtual worker injured in a work accident, obtained by combining the selected data and by creating an optimal rehabilitation program applicable to the virtual worker injured in an industrial accident. However, the example cases do not necessarily have to include a virtual case. If there are enough actual cases based on the number of attributes stored in a case base table, virtual cases can be excluded.

In 13 For example, the rehabilitation program table 213, case base table 211, and rehabilitation detail table 212 stored in memory 210 are represented as an entity relationship diagram (ERD). Regarding 13 the rehabilitation details table 212 is a table that links the rehabilitation program table 213 and the case base table 211 . The rehabilitation details table 212 has two foreign keys. One is a register number obtained from the case base table 211, the other is a program number obtained from the rehabilitation program table 213. The program number is a unique value for locating a specific rehabilitation program. In addition to the two foreign keys, the rehabilitation details table 212 stores data generated while the rehabilitation program table 213 and the case base table 211 are related.

It will come back on at this point 10 returned, in which the input unit 220 receives information and/or a command from the job advisor. For example, the input unit 20 receives attribute information related to the worker injured in a work accident. As attribute information, the 16 in 12 received parts of the attribute information shown. In addition, the input unit 220 receives an instruction to execute a recommendation for a set of rehabilitation programs.

The output unit 230 outputs a result of command processing. For example, when a set of rehabilitation programs suitable for the worker injured in a work accident is found, the output unit 230 outputs the set of rehabilitation programs found. For this purpose, the output unit 230 can include a display 231 for outputting a video signal and a loudspeaker 232 for outputting an audio signal.

The search unit 250 searches the case base tables 211 stored in the memory 210 for a similar worker who suffered an accident similar to the worker injured in a work accident. For this purpose, the search unit 250 performs the search using an “injury and disease group code” among the attribute information received by the input unit 220 . A search result is passed to the similarity calculator 260, which is described below.

The similarity calculator 260 extracts the 15 attributes other than the injury and disease group code from the case base table of the similar industrially injured worker found by the search unit 250, and calculates a similarity to the industrially injured worker using the extracted attributes. According to the exemplary embodiment, the similarity calculator 260 comprises a first similarity calculator 261, a second similarity calculator 262 and a summation unit 263.

The first similarity calculator 261 calculates a similarity between the industrially injured worker and the similar industrially injured worker based on continuous attributes (ie, numerical data) among the in 12 displayed attributes. Here a Euclidean distance can be used to calculate the similarity. If two points p and q have coordinates p = (p1, p2, ...., pn) and q = (q1, q2, ...., qn) in a multidimensional space, the Euclidean distance between the two points becomes calculated theoretically as in Equation 1 below.

$$\sqrt{{\left({\mathrm{<figure-callout\; id="1"\; label="q"\; filenames="DE102022105049A1\_0003.png,DE102022105049A1\_0006.png"\; state="\{\{state\}\}">q</figure-callout>}}_1-p_1\right)}^2+{\left({\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="DE102022105049A1\_0006.png,DE102022105049A1\_0007.png"\; state="\{\{state\}\}">q</figure-callout>}}_2-p_2\right)}^2+\cdots +{\left(q_n-p_n\right)}^2}=\sqrt{{\displaystyle \sum _{i=1}^n{\left(q_i-p_i\right)}^2}}$$

The second similarity calculator 262 calculates a similarity between the industrially injured worker and the similar industrially injured worker based on nominal attributes among the in 12 displayed attributes. Here a Hamming distance can be used to calculate the similarity. The Hamming distance is used in a method of measuring the similarity between two nominal variables, using the number of characters present at the same position in two strings of the same length but different from each other. In other words, it compares how many characters differ in either of the two character strings between which a similarity is calculated to make the character strings equal. A Hamming distance is calculated as in Equation 2 below.

$$\begin{array}{l}{D}_H={\displaystyle \sum _{i=1}^k\left|x_i-y_i\right|}\\ x=y\Rightarrow D=0\\ x\ne y\Rightarrow D=1\end{array}$$

The summation unit 263 calculates the final similarity by adding the Euclidean distance value and the Hamming distance value. For example, the summation unit 263 calculates a weighted sum of the attribute-specific Euclidean distance values and a weighted sum of the attribute-specific Hamming distance values, and then determines a sum of the two weighted sums as a final similarity. As another example, the summing unit 263 calculates a grand total of the attribute-specific Euclidean distance values and a grand total of the attribute-specific Hamming distance values, and then calculates a weighted sum of the calculated sum of the Euclidean distance values and the calculated sum of the Hamming distance values. The calculated weighted sum is then determined as the final similarity. In other words, summation unit 263 may give each attribute a weight regardless of whether the attribute is a continuous attribute or a nominal attribute, or give an attribute a weight depending on whether the attribute is a continuous attribute or a nominal attribute. A method for the final similarity calculation of the summation unit 263 is presented as equations in 14 shown.

At this point will turn on again 10 in which the similarity calculator 260 calculates the final similarities with all similar workers found by the search unit 250. The similarity calculator 260 then supplies the calculated final similarities to the control unit 240.

The control unit 240 connects and controls the components. The control unit 240 selects a similar worker injured in a work accident with the greatest final similarity among the similar workers found by the search unit 250 . Then configure the Control unit 240 displays a recommendation summary containing a set of rehabilitation programs provided to the selected similar worker. The configured recommendation overview is in 15A shown.

Regarding 15A A recommendation overview 700 includes a recommendation result display area 710, a change button 720 and a save button 730.

The field-specific rehabilitation programs are arranged on a vertical axis of the recommendation results display area 710 . In particular, the rehabilitation programs from the fields of medicine, social psychology and work are arranged in a vertical sequence. Medical care grades are sequentially arranged in a horizontal direction on a horizontal axis of the recommendation result display area 710 . Medical care stages are divided into an acute phase, a recovery phase, a stable phase, a full recovery phase and a return to work phase.

As in the recommendation result display area 710 in 15A represented, rehabilitation programs included in a recommended set of rehabilitation programs are with a different color, e.g. B. gray, highlighted as rehabilitation programs that are not included in the recommended set of rehabilitation programs. Also, the rehabilitation programs included in a recommended set of rehabilitation programs are displayed according to the medical care levels. When the rehabilitation programs are displayed in this way, the job counselor can intuitively know what type of rehabilitation program is suitable for each level of medical care, in contrast to a case where the rehabilitation programs included in a recommended set of rehabilitation programs are only in the text are listed.

The work advisor can then recognize the content of the recommendation result display area 710 and set the displayed content as the rehabilitation plan for the worker injured in a work accident as recommended. When the rehabilitation plan is set, the job counselor selects the save button 730. Then, the set of rehabilitation programs displayed in the recommendation result display area 710 is saved in the memory 210 as a new case.

However, for example, the work advisor can also recognize the content of the recommendation result display area 710, change a type of rehabilitation program to be provided to the worker injured in a work accident and/or a medical care level for the provision of the rehabilitation program in the displayed content, and then set the changed content as a rehabilitation plan for the worker injured in a work accident. When the content of the recommendation result display area 710 is to be changed, the job advisor selects the change button 720 so that a selection area 711 of the recommendation result display area 710 can be activated. Then, from among the cells arranged in the selection area 711, the job advisor selects a cell corresponding to a type of rehabilitation program to be offered to the worker injured in a work accident and/or a medical care level for providing the rehabilitation program. The cell selected by the work advisor is highlighted compared to unselected cells. The work counselor then selects the save button 730. The set of rehabilitation programs displayed in the selection area 711 is then stored in the memory 210 as a new case.

When a rehabilitation plan for the worker injured in a work accident is prepared through such a process, the labor counselor provides the worker injured in a work accident with a rehabilitation program according to the prepared rehabilitation plan. Every time the worker completes the rehabilitation program, the work advisor stores data about the completed rehabilitation program in the recommendation device for rehabilitation program sets 200. Data generated every time the worker completes a rehabilitation program can be stored in the rehabilitation details table that regarding 13 is described.

In some cases, during the process of providing a rehabilitation program based on the rehabilitation plan for the worker injured in a work accident, it may be necessary to change the types of rehabilitation programs according to the condition of the worker injured in a work accident. In this case, the job advisor can search for a rehabilitation program set suitable for the worker injured in a work accident in the rehabilitation program set recommending device 200 . When a set of rehabilitation programs found is displayed on the display 231, the job counselor can select the change button 720 and change the information.

A case of selecting a similar worker injured in a work accident with the greatest ultimate similarity among similar workers found to configure the recommendation map 700 was described above as an example. According to another exemplary embodiment, similar workers injured in a work accident with a final similarity exceeding a certain reference may be selected and recommendation summaries configured separately for the selected similar workers. In this case, the configured recommendation overviews can be displayed in time intervals one after the other via the display 231 or can be displayed simultaneously via divided areas of the display 231 . Alternatively, as in 15B As shown, a plurality of icons 740 corresponding to a plurality of recommendation summaries are displayed on the display 231.

As in 15B As illustrated, the plurality of icons 740 may be arranged in a row from left to right, and an icon corresponding to a recommendation summary of a similar worker injured in a work accident with a greater final likeness may be positioned further to the left. Additionally, an icon corresponding to a recommendation summary 700 currently displayed on the display 231 may be highlighted relative to the other icons. In 15B For example, the leftmost of the three icons 740 displayed in the recommendation map 700 is highlighted, and thus the currently displayed recommendation map 700 is a recommendation map of a similar worker with the greatest ultimate similarity among the previously selected similar workers. In this case, if the job counselor selects a particular non-highlighted icon, the recommendation map 700 displayed via the display 231 may be switched to a recommendation map (not shown) corresponding to the selected icon. According to such an exemplary embodiment, when an icon is clicked, a recommendation summary corresponding to a similar worker injured in a work accident with the icon is displayed, and thus a job counselor can easily compare the rehabilitation program set information of several similar workers.

16 12 is a flowchart showing a method for recommending a set of rehabilitation programs for a worker injured in a work accident according to an exemplary embodiment of the present invention.

First, the rehabilitation program set recommending device 200 obtains attribute information on a worker injured in a work accident from a work counselor (S810). In other words, 16 pieces of attribute information including an injury and disease code are entered, as in 12 shown.

Then, similar workers who have suffered accidents similar to those of the worker injured in a work accident are searched for using the injury and disease group code of the worker injured in a work accident among the received attribute information (S820). In other words, among the case base tables stored in the memory 210, case base tables containing an injury and illness group code identical to the injury and illness group code of the worker injured in a work accident are searched.

Then, a similarity between the worker injured in a work accident and each of the similar workers being searched is calculated (S830). The method S830 includes a process of calculating the Euclidean distance value of the continuous attributes, a process of calculating the Hamming distance value of the nominal attributes, a process of calculating a weighted sum of the Euclidean distance value and the Hamming distance value, and a process of determining the calculated weights Sum as final similarity.

When a resemblance to each of the searched similar injured workers is calculated (S830), a similar injured worker having the greatest resemblance is selected (S840). The method S840 includes a process of selecting a similar injured worker with the greatest ultimate similarity.

Thereafter, a recommendation map is configured to include a set of rehabilitation programs provided to the selected similar injured worker (S850). In the recommendation overview, the rehabilitation programs included in the set of rehabilitation programs can be displayed according to the medical care levels.

The configured recommendation overview is then output (S860). The method S860 includes the act of displaying the recommendation summary on the display 231. The job counselor can then recognize the content of the displayed recommendation summary and then change the content as needed.

It is then determined whether a store command should be entered. If fixed is set that no save command should be entered, the displayed recommendation summary is retained unchanged. When it is determined that a save command should be entered, the content displayed on the recommendation map is saved in memory 210 as a new case.

A system and method for recommending rehabilitation program sets according to exemplary embodiments of the present invention have been described above. In the exemplary embodiment, a case where the rehabilitation program set recommending system 1 includes the big data platform 100 and the rehabilitation program set recommending device 200 was taken as an example.

Although not shown in the drawing, a rehabilitation program set recommendation system according to another exemplary embodiment may include a big data platform, a rehabilitation program set recommendation device, and a terminal device.

The terminal is an electronic device used by an employment counselor. The electronic device includes a wired or wireless communication device that performs a computing operation. An example of a wired communication device can be a computer (PC). Examples of wireless communication devices may include second-generation (Code-Division Multiple Access (CDMA) and Global System for Mobile Communications (GSM)) mobile terminals, third-generation (CDMA2000, Wideband Code-Division Multiple Access (W-CDMA) mobile terminals), time Division Synchronous CDMA (TD-SCDMA) and Universal Mobile Telecommunication System (UMTS)), the 3.5th generation (High Speed Downlink Packet Access (HSDPA)), the 3.75th generation (High Speed Uplink Packet Access (HSUPA)), the 3.9 th generation (Evolved High Speed Packet Access (HSPA+) and Long Term Evolution (LTE)), fourth generation (LTE-advanced), fifth generation (5G), etc.

The terminal receives attribute information of the worker injured in a work accident from the job adviser and transmits the received attribute information to the rehabilitation program set recommending device. The terminal may also receive and display a recommendation summary generated by the rehabilitation program set recommender.

The terminal includes an input unit for receiving various information and/or various commands from the job advisor, a communicator for communicating with the rehabilitation program set recommending device, and an output unit for outputting recommendation information received from the rehabilitation program set recommending device. The output unit can include a display for outputting a video signal and a loudspeaker for outputting an audio signal.

The rehabilitation program recommending device may additionally include a communicator for communicating with the terminal.

Exemplary embodiments of the present invention have been described above. In addition to the exemplary embodiments described above, embodiments of the present invention may be implemented by a medium, for example a computer-readable medium, containing computer-readable code or instructions for controlling at least one processing element of the embodiments described above. The medium can be a medium or media that enable the storage and/or transmission of the computer-readable code.

The computer-readable code can be not only recorded on a medium but also transmitted over the Internet. The medium can be, for example, a storage medium such as a magnetic storage medium (e.g. read only memory (ROM), floppy disk, hard disk, etc.) or an optical storage medium (e.g. compact disc (CD)-ROM, a Blu-Ray, a Digital Versatile Disc (DVD), etc.) or a transmission medium such as carrier waves. Because the media can be provided over a distributed network, the computer readable code can be stored or transmitted and executed in a distributed manner. Further, the processing element may comprise a processor or a computer processor, for example, and may be distributed and/or contained in one device.

According to the present invention, since a set of rehabilitation programs is recommended that is expected to be suitable for a worker injured in a work accident, a work consultant using the recommended set of rehabilitation programs can easily make a rehabilitation plan for the worker injured in a work accident create.

According to the present invention, even an inexperienced job counselor can easily make a rehabilitation plan for a worker injured in a work accident.

According to the present earth invention, a job counselor can make a rehabilitation plan as well without creating electronic health records of a worker injured in a workplace accident.

Although embodiments of the present invention have been described above with reference to the accompanying drawings, those skilled in the art should understand that the present invention can be embodied in other specific forms without departing from the technical spirit or essential characteristics of the present invention. It should therefore be understood that the embodiments described above are in all respects exemplary and not restrictive.

QUOTES INCLUDED IN DESCRIPTION

This list of documents cited by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent Literature Cited

• KR 1020210030098 [0001]

Claims (10)

Apparatus for recommending a set of rehabilitation programs for a worker injured in an industrial accident, the apparatus comprising: a memory configured to store case base tables including data on workers injured in a work accident who received rehabilitation service according to a set of rehabilitation programs and successfully returned to work; a search unit configured, after inputting attribute information relating to a worker injured in a work accident, and using the inputted attribute information to search the memory and search for similar workers who have suffered accidents similar to the worker injured in a work accident ; a similarity calculator configured to calculate a similarity between the worker injured in a work accident and each of the searched workers who suffered similar work accidents; a control unit configured to select an injured worker with the greatest calculated similarity and to generate a recommendation map displaying the rehabilitation programs included in a set of rehabilitation programs and provided to the selected worker with similar attributes according to phases of medical care; and an output device configured to output the recommendation summary. device after claim 1 , where the case base tables contain continuous attributes and nominal attributes. device after claim 2 , wherein the continuous attributes include at least one characteristic of a severity index, age, average wage, number of days worked, number of workers injured at a workplace, and total score on a multidimensional psychological test. device after claim 2 , wherein the nominal attributes include at least one of an injury and illness group code, an accident type code, an employee status code, an industry code, an occupation code, an indication of whether a workplace has been notified of a corresponding accident, an indication of whether a corresponding employee work at the same workplace would like, an indication of whether the relevant employee is married, a counseling target trend code and an indication of whether a relevant employee should be recommended a job. device after claim 2 , wherein the similarity calculator comprises: a first similarity calculator configured to calculate a Euclidean distance value according to each of the continuous attributes; a second similarity calculator configured to calculate a Hamming distance value according to each of the nominal attributes; and a summation unit configured to calculate a weighted sum of the Euclidean distance values and the Hamming distance values. device after claim 5 , where the summation unit weights the Euclidean distance values using weight factors calculated separately on the continuous attributes and the Hamming distance values using weight factors calculated separately on the nominal attributes. device after claim 1 , wherein the inputted attribute information includes at least an injured worker severity index, an injured worker's age, an injured worker's average wage, a number of days worked by the injured worker, a number of workers who have suffered an industrial injury at a workplace of the injured worker, a The injured worker's total score on a multidimensional psychological test, the injured worker's injury and illness group code, an injured worker's accident type code, an injured worker's worker status code, an injured worker's branch code, an injured worker's occupation code, an indication of whether a job has a was notified of the relevant accident, an indication of whether the injured worker would like to work at the same job, an indication of whether the injured worker is married, a counseling target trend code, and an indication of whether the injured worker a job should be recommended to the last worker. device after claim 7 wherein the search unit searches for case base tables that have the same injury and disease group code as the injury and disease group code. device after claim 1 , the case base tables in a mass data platform created by merging a labor insurance system of the Korea Workers' Welfare and Compensation Bureau (Korea Workers' Wel fare and Compensation Service) is compiled with a medical information system of hospitals, stored and managed. device after claim 9 wherein the big data platform comprises: a star schema in which the case base table is considered a fact table and tables selected in the employment insurance system are arranged around the case base table as dimension tables; a star schema in which two main tables selected in the medical information system are regarded as separate fact tables, and tables other than the two main tables are arranged around the main tables as dimension tables; and a link table configured to link the case base table to the two main tables, the link table containing as main attributes a register number received from the case base table and a personal identification number received from the two main tables and an accident date.

Images