Programs, information processing devices, methods, and systems

Abstract

The aim is to improve the accuracy of predicting employee turnover risk. [Solution] A program according to one aspect of the present disclosure is a retirement prediction program for operating a computer comprising a processor and memory, wherein the program causes the processor to perform the following steps: acquire quantitative data relating to an employee's work attitudes; acquire quantitative score data converted from qualitative data relating to an employee's work attitudes using a generating AI; take the quantitative data and score data as input and use a retirement prediction model constructed based on data of past retirees to calculate the risk that an employee will retire within a predetermined period; and output the retirement risk to a terminal device used by an administrator.

Description

Technical Field

[0001] The present disclosure relates to a program, an information processing apparatus, a method, and a system.

Background Art

[0002] For a company, retaining employees is an important management issue, and there is a technology for detecting signs of employee resignation at an early stage.

[0003] Patent Document 1 discloses a technology for analyzing voice communication data such as interviews to predict future actions of employees. There are also application programs for regularly measuring the state of employees, such as engagement surveys and employee satisfaction surveys.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, with the conventional method of analyzing communication data as disclosed in Patent Document 1, it was not always possible to sufficiently draw out the true feelings of employees and accurately capture the fundamental factors leading to a high risk of resignation. Therefore, a new mechanism capable of accurately predicting the resignation risk has been demanded.

[0006] Based on such recognition of the problem, the present disclosure aims to improve the accuracy prediction of the resignation risk.

Means for Solving the Problems

[0007] To solve the above problems, a program according to one aspect of the present disclosure is a retirement prediction program for operating a computer comprising a processor and memory, wherein the program causes the processor to perform the following steps: acquire quantitative data on an employee's work attitudes; acquire quantitative score data converted from qualitative data on an employee's work attitudes using a generating AI; take the quantitative data and score data as input and use a retirement prediction model constructed based on data of past retirees to calculate the risk that an employee will retire within a predetermined period; and output the retirement risk to a terminal device used by an administrator. [Effects of the Invention]

[0008] According to this disclosure, it is possible to improve the accuracy of predicting employee turnover risk. [Brief explanation of the drawing]

[0009] [Figure 1] This block diagram shows the overall configuration of a retirement prediction system according to one embodiment of the present disclosure. [Figure 2] This block diagram shows a functional configuration example of a terminal device according to this embodiment. [Figure 3] This block diagram shows an example of the functional configuration of the server according to this embodiment. [Figure 4] This figure shows an example of the data structure of the employee data table used in this embodiment. [Figure 5] This figure shows an example of the data structure of the analysis results table used in this embodiment. [Figure 6] This flowchart shows an example of the retirement prediction processing flow according to this embodiment. [Figure 7] This figure shows an example of a retirement risk analysis screen displayed on the terminal device according to this embodiment. [Figure 8] This figure shows an example of interview support information displayed on the terminal device according to this embodiment. [Figure 9] This is a block diagram representing the basic hardware configuration of a computer. [Modes for carrying out the invention]

[0010] The embodiments of this disclosure will be described below with reference to the drawings. In all the drawings illustrating the embodiments, common components are denoted by the same reference numerals, and repeated explanations are omitted. The following embodiments are not intended to unduly limit the content of this disclosure as described in the claims. Not all components shown in the embodiments are necessarily essential components of this disclosure. Also, each drawing is a schematic diagram and is not necessarily a strict illustration.

[0011] Furthermore, in the following description, "processor" refers to one or more processors. At least one processor is typically a microprocessor such as a CPU (Central Processing Unit), but may be another type of processor such as a GPU (Graphics Processing Unit). At least one processor may be single-core or multi-core.

[0012] Furthermore, at least one processor may be a broad-sense processor, such as a hardware circuit that performs some or all of the processing (e.g., an FPGA (Field-Programmable Gate Array) or an ASIC (Application Specific Integrated Circuit)).

[0013] Furthermore, in the following explanation, we may use expressions such as "xxx table" to describe information that yields an output for a given input. This information can be data with any structure, or it can be a learning model such as a neural network that generates an output for a given input. Therefore, "xxx table" can be referred to as "xxx information."

[0014] In the following description, the configuration of each table is an example. One table may be divided into two or more tables, or all or part of two or more tables may be combined into one table.

[0015] In the following description, the "program" may be used as the subject to describe the process. However, since the program is executed by a processor to perform the defined process while appropriately using a storage unit and / or an interface unit, etc., the subject of the process may be the processor (or a device such as a controller having the processor).

[0016] The program may be installed in a device such as a computer, or may be in, for example, a program distribution server or a computer-readable (e.g., non-transitory) recording medium. Also, in the following description, two or more programs may be realized as one program, or one program may be realized as two or more programs.

[0017] In the following description, an identification number is used as the identification information for various objects, but identification information of other types (e.g., an identifier including letters or symbols) may be adopted.

[0018] In the following description, when explaining without distinguishing between elements of the same type, reference signs (or common signs among the reference signs) are used, and when explaining while distinguishing between elements of the same type, the identification numbers (or reference signs) of the elements may be used.

[0019] In the following description, the control lines and information lines indicate those considered necessary for the description, and not necessarily all control lines and information lines on the product are shown. All components may be interconnected.

[0020] Each information processing device consists of a computer equipped with an arithmetic unit and a memory device. The basic hardware configuration of the computer and the basic functional configuration of the computer realized by said hardware configuration will be described later. For each of the terminal device 10 and the server 20, explanations that overlap with the basic hardware configuration and basic functional configuration of the computer described later will be omitted.

[0021] <1. Overview> The employee turnover prediction system 1 is a service that predicts employee turnover risk by integrating quantitative and qualitative data and recommends specific measures that managers should take. In this system, server 20 acquires quantitative data such as employee surveys. Simultaneously, conversational AI system 40 acquires quantitative score data converted from employee dialogue logs (qualitative data) by a generating AI. The acquired quantitative data and score data are then input into the employee turnover prediction model to calculate the turnover risk for each employee. This calculation result is displayed on terminal device 10 operated by the manager.

[0022] In this specification, "qualitative data" refers to text-based data that includes the content of employees expressing their concerns, opinions, or feelings regarding their work attitudes. This qualitative data serves as the source for converting score data to be input into the employee turnover prediction model, and the most typical example is the dialogue log with the conversational AI system 40 in this embodiment. However, the present invention is not limited to this form, and various text data that express employees' true feelings and psychological states, such as minutes of 1-on-1 meetings between superiors and subordinates, records of interviews with HR personnel, or free-response answers in employee satisfaction surveys, can also be included in the "qualitative data" targeted by the present invention. Here, "score data" refers to quantitative evaluation values ​​obtained by the generating AI (conversational AI system 40) analyzing the content of qualitative data such as dialogue logs based on predetermined evaluation criteria, such as "job satisfaction" or "interpersonal relationships." This makes it possible to treat text information, which is difficult to analyze as is, as objective numerical values.

[0023] <2. System Configuration> Figure 1 is a block diagram showing an example of the overall configuration of the employee turnover prediction system 1 related to this disclosure. In this disclosure, we will explain using as an example a case in which the employee turnover prediction system 1 is introduced within a corporate organization, and the administrator of that organization uses the system as a user to analyze and manage the risk of employee turnover.

[0024] As shown in Figure 1, this embodiment consists of, for example, a terminal device 10 used by an administrator, a server 20 that performs various data processing, and a generation AI system 30 that is responsible for calculating employee turnover risk and generating recommended actions. System 1 is connected via a network 80 to an external system, an interactive AI system 40, which is used to acquire employee dialogue logs, and a talent management system 50, which is a source of quantitative employee data and also serves as a destination for sharing analysis results. In Figure 1, the interactive AI system 40 and the talent management system 50 are shown as separate systems, but they may be integrated into a single system.

[0025] Figure 1 shows an example where System 1 includes one terminal device 10, for the sake of illustration simplification. However, in reality, System 1 may include multiple terminal devices 10 for use by multiple administrators.

[0026] Figure 1 shows an example where the generative AI system 30 is independent of the server 20, but the server 20 may include the functions of the generative AI system 30. In that case, the server 20 stores the large-scale language model (LLM) used for data transformation in its own memory and completes the processing internally.

[0027] Terminal device 10 is an information processing device operated by a manager, such as an employee's supervisor or a human resources officer. In this specification, "manager" is not limited to a specific job title, but refers to any person who is responsible for employee evaluation, training, labor management, etc., and who is in a position to take action to prevent employee turnover. Specifically, this could include an employee's direct supervisor, a department head such as a group leader or manager, a human resources officer, or management. The manager accesses the employee turnover prediction system 1 via a dedicated application or web browser installed on this terminal device 10 (e.g., a PC, laptop PC, tablet, etc.) to view analysis results or input interview results, etc.

[0028] As shown in Figure 1, the terminal device 10 includes a communication interface 12, an input device 13, an output device 14, memory 15, storage 16, and a processor 19. The input device 13 is a keyboard or mouse, etc., for receiving input operations from the administrator. The output device 14 is a display, etc., for displaying analysis results, etc.

[0029] Server 20 is an information processing device that plays a central role in this system. In response to requests from terminal devices 10, Server 20 integrates data stored in its own memory or data acquired from external systems, performs analysis processing using machine learning models, and transmits the results to terminal devices 10.

[0030] As shown in Figure 1, the server 20 includes a communication interface 22, an input / output interface 23, memory 25, storage 26, and a processor 29. The memory 25 and storage 26, which function as storage units, store programs or various data for providing this system. This data includes personnel data that manages basic employee information or performance history, or employee survey response data.

[0031] The generation AI system 30 receives the results of the employee turnover risk analysis (risk factors, etc.) from the server 20 and performs advanced analytical processing, such as generating recommended actions for administrators, based on the instructions.

[0032] The generation AI system 30 is, for example, a cloud server that has an LLM. The number of LLMs included in the generation AI system 30 may be one or multiple.

[0033] LLM (Language Modeling) is a single-modal natural language model built by training on large amounts of text data, and is used for many NLG (Natural Language Generation) tasks, such as generating responses to specific questions, automatically generating sentences, and summarizing text. LLM is an example of a generative AI model. Examples of LLMs include the following: • OpenAI: GPT-4 Google: Gemini 1.5 Flash Anthropic:Claude 3.5 Sonnet

[0034] In this embodiment, the "execution of the retirement prediction model" and the "generation of recommended actions" are primarily described as being handled by the generation AI system 30. However, the present invention is not limited to this, and the server 20 may be responsible for executing the retirement prediction model. The server 20 is mainly responsible for data linkage with external systems, issuing instructions to the generation AI system 30, and managing and presenting analysis results.

[0035] In the "Execution of the Retirement Prediction Model," the server 20 first obtains "quantitative data" and "score data" of the employees to be analyzed from an external system, etc., and sends them to the generating AI system 30. Next, the generating AI system 30, having received the instructions, uses this received data as input and calculates the retirement risk using a predefined retirement prediction model (for example, a machine learning model or statistical analysis model described later).

[0036] The generation AI system 30, upon receiving instructions, interprets the complex relationships within the data and generates the rationale for the calculated turnover risk in natural language. The advantage of using generation AI is that it not only calculates a score, but also presents the rationale for "why this risk score was reached" in natural language that is easy for humans to understand. For example, it can generate a specific explanation such as, "Recent conversations have shown an increase in statements like 'I can't get concrete advice even when I consult with my supervisor' and 'I can't see a clear future for my career,' suggesting that dissatisfaction with the lack of career support may be increasing." This prevents the prediction results from becoming a black box and increases the manager's sense of satisfaction.

[0037] In "Generating Recommended Actions," Server 20 transmits the "Resignation Risk Factors" identified through analysis, the "Personality Trait Data" stored in Employee Data Table 2021, and the "Success Stories Data" from past interventions to the AI ​​Generation System 30 as structured data. The instructions given at that time would be something like, "Based on this employee's risk factors and personality traits, propose an effective interview process that the manager should follow, along with specific questions and realistic next actions that take into account the constraints of the employee's environment."

[0038] The AI ​​generation system 30, upon receiving instructions, leverages its advanced language generation capabilities to generate personalized recommendations for individual employees, rather than simply using canned responses. For example, it suggests fact-checking-focused dialogues for employees with logical thinking skills, and emotionally resonant language for employees who value empathy. Furthermore, for identified risk factors (e.g., "lack of growth opportunities"), it generates specific questions and next actions that lead to problem-solving, such as "What kind of work would you like to take on in the future?" and "What kind of support can the company provide to help you achieve that?"

[0039] The engagement status, detailed factors, and recommended actions for each individual, generated in this way, are clearly displayed on the terminal screens of administrators and interviewers as interview support information, as shown in Figure 8.

[0040] The conversational AI system 40 is an existing system located outside of system 1 that provides a conversational service with an AI mentor, etc., that employees can use at their discretion, and manages the conversation logs. This conversational service includes, for example, a chatbot that listens to employees' work-related concerns. Server 20 obtains "qualitative data," which is the employee's conversation log, and "score data," which is generated based on it, from this system via network 80.

[0041] Talent management system 50 is an existing system located outside of system 1 that centrally manages employee personnel information, skills, career plans, etc. Server 20 can link the results of the analysis of employee turnover risk or risk factors performed by this system with talent management system 50 and utilize them for employee management.

[0042] Each information processing device or system, such as the terminal device 10, server 20, generation AI system 30, conversational AI system 40, and talent management system 50, may be composed of a computer 90 equipped with a computing device and a memory device. The basic hardware configuration of the computer 90 and the basic functional configuration of the computer 90 realized by said basic hardware configuration will be described later. Note that explanations of each component that overlap with the basic hardware configuration of the computer 90 and the basic functional configuration of the computer will be omitted.

[0043] <3. Configuration of terminal equipment> Figure 2 is a block diagram showing an example of the functional configuration of the terminal device 10. As shown in Figure 2, the terminal device 10 comprises a communication unit 120, an input device 13, an output device 14, an optional voice processing unit 17, a microphone 171, a speaker 172, a camera 160, a location information sensor 150, a storage unit 180, and a control unit 190. Each block included in the terminal device 10 is electrically connected, for example, by a bus. In this embodiment, the terminal device 10 primarily provides an interface for the user, who is an administrator, to use the retirement prediction service.

[0044] The communication unit 120 performs processing such as modulation and demodulation processing for the terminal device 10 to communicate with other devices. The communication unit 120 performs transmission processing on the signal generated by the control unit 190 and sends it to an external source (for example, the server 20). The communication unit 120 performs reception processing on the signal received from the external source and outputs it to the control unit 190. As a result, the analysis request or action record entered by the administrator is sent to the server 20, and the terminal device 10 receives employee turnover risk, risk factors, recommended actions, etc. from the server 20.

[0045] The input device 13 is a device for a user (administrator) operating the terminal device 10 to input instructions or information. The input device 13 can be implemented, for example, by a touch-sensitive device 131 on which instructions are input by touching the operating surface. If the terminal device 10 is a PC, the input device 13 may be implemented by a keyboard, mouse, etc. The input device 13 converts the instructions input by the user (for example, specifying the employee to be analyzed, inputting the results after performing the recommended action, etc.) into an electrical signal and outputs the electrical signal to the control unit 190.

[0046] The output device 14 is a device for presenting information to the user (administrator) operating the terminal device 10. The output device 14 is implemented, for example, by a display 141. The display 141 displays data (for example, employee turnover risk, turnover risk factors, recommended actions, etc.) according to the control of the control unit 190. The display 141 is implemented, for example, by an LCD (Liquid Crystal Display) or an organic EL (Electro-Luminescence) display.

[0047] The audio processing unit 17 performs, for example, digital-to-analog conversion processing of the audio signal. The microphone 171 receives an audio input and provides the audio signal corresponding to the audio input to the audio processing unit 17. The speaker 172 converts the audio signal provided by the audio processing unit 17 into audio and outputs the audio to the outside of the terminal device 10. In this embodiment, these audio-related components can be used when an administrator inputs interview records by voice or receives notifications from the system by voice.

[0048] Camera 160 is a device that receives light using a photodetector and outputs it as a shooting signal. In the retirement prediction service of this embodiment, the camera function is not essential, but it can be used, for example, if a video call function is envisioned for managers to conduct interviews with employees remotely.

[0049] The location information sensor 150 is a sensor that detects the location of the terminal device 10, and is, for example, a GPS (Global Positioning System) module. In the retirement prediction service of this embodiment, the location information sensor is not essential, but it can be used, for example, when displaying an analysis report of employees according to the administrator's location.

[0050] The storage unit 180 stores data and programs used by the terminal device 10. For example, the storage unit 180 stores user information 181, application programs for using the retirement prediction service, or configuration information.

[0051] User information 181 includes, for example, information about the user (administrator) who uses terminal device 10. User information includes, for example, the user's name, login ID, administrative department, contact information, service usage history, etc.

[0052] The control unit 190 is realized when the processor reads a program stored in the memory unit 180 and executes instructions contained in the program. The control unit 190 controls the operation of the terminal device 10. By operating according to the program, the control unit 190 performs the functions of an operation reception unit 191, a transmission / reception unit 192, and a presentation control unit 193.

[0053] The operation reception unit 191 processes instructions or information input from the input device 13. Specifically, for example, the operation reception unit 191 can receive information such as the selection of an employee to be analyzed or text input of interview results, which is input from a touch-sensitive device 131 or the like. The operation reception unit 191 can also receive voice instructions or voice memos input from the microphone 171.

[0054] The transmitting / receiving unit 192 performs processing to enable the terminal device 10 to send and receive data with external devices such as the server 20 in accordance with a communication protocol. Specifically, for example, the transmitting / receiving unit 192 sends analysis requests or action records entered by the administrator to the server 20. The transmitting / receiving unit 192 also receives information provided by the server 20 (such as employee turnover risk, risk factors, and recommended actions).

[0055] The presentation control unit 193 controls the output device 14 to present information provided by the server 20 to the user (administrator). Specifically, for example, the presentation control unit 193 displays various analysis information regarding employee turnover risk transmitted from the server 20 on the display 141. The presentation control unit 193 can also output notifications from the server 20 as audio from the speaker 172.

[0056] <4. Server Configuration> Figure 3 is a block diagram showing an example of the functional configuration of the server 20 shown in Figure 1. As shown in Figure 3, the server 20 performs the functions of a communication unit 201, a storage unit 202, and a control unit 203.

[0057] The communication unit 201 performs processing to enable the server 20 to communicate with external devices (terminal device 10, generation AI system 30, conversational AI system 40, talent management system 50).

[0058] The memory unit 202, implemented by memory 25 and storage 26, stores data and programs used by the server 20 to provide employee retirement prediction services. The memory unit 202 stores the employee data table 2021 and the analysis results table 2022 as primary data. The employee data table 2021 stores quantitative data, score data, and personality trait data of employees that form the basis of the analysis. The analysis results table 2022 stores analysis results such as retirement risk calculated using this data, identified risk factors, and generated recommended actions. It also stores machine learning models used for risk calculation, or various settings (thresholds, weights, etc.) used for analysis.

[0059] The control unit 203 is implemented by the processor 29 executing a program stored in the memory unit 202, and controls the operation of the entire server 20. The control unit 203 can perform the functions of a data acquisition module 2031, a risk prediction module 2032, an action recommendation module 2033, and an effectiveness measurement module 2034.

[0060] The data acquisition module 2031 collects the data necessary for analysis. Specifically, it acquires quantitative data such as personnel data and survey response results from an external talent management system 50. It also acquires qualitative data such as employee dialogue logs and quantitative score data obtained by the AI ​​generating the qualitative data from the conversational AI system 40 via the communication unit 201, and stores them in the employee data table 2021.

[0061] The risk prediction module 2032 uses the data acquired by the data acquisition module 2031 to instruct the generation AI system 30 to analyze employee turnover risk. Specifically, it sends the data necessary for the analysis to the generation AI system 30 as a prompt and receives the processing results (calculated employee turnover risk, risk factors, etc.). Based on these instructions, the generation AI system 30 executes an employee turnover prediction model, such as a statistical analysis model or a machine learning model, as described later. Once processing is complete, the risk prediction module 2032 receives the analysis results, such as the calculated employee turnover risk and the underlying risk factors, as structured data (e.g., in JSON format).

[0062] In the statistical test used in this embodiment, the turnover rate of the group of employees who answered "dissatisfied" to a specific survey item is first compared with the average turnover rate of all employees. Next, the difference in turnover rates is scientifically proven, for example, using a statistical method such as a binomial test, to be not due to chance, and the correlation between the item and turnover is objectively determined. The "turnover risk factors" identified in this way are used for two different purposes. The first use is to select explanatory variables to improve the accuracy and interpretability of the machine learning model, and only items deemed statistically significant are adopted as the main explanatory variables of the machine learning model. The second use is to make the statistical testing method itself function as an independent "statistical analysis model" for determining turnover risk. This makes it possible to perform risk assessment as a simpler model with higher interpretability, separate from the complex machine learning model. Note that the results of risk assessment by this statistical analysis model are not limited to being output as a specific turnover probability (%), but may also be determined as a category such as "high risk" or "caution" based on whether the identified risk factors exceeded a predetermined threshold.

[0063] The machine learning model used in this embodiment is constructed by training it with data from several hundred past former and current employees as training data. In this training process, the model learns the correlations between which combinations or trends of data items (employee attributes, survey responses, scores obtained from dialogue logs, etc.) are likely to lead to "employee resignation." Specifically, it calculates a coefficient representing the degree to which each data item influences the probability of resignation.

[0064] As an example of a model, the logistic regression model, a highly interpretable statistical analysis model, can be used. The logistic regression model is a statistical method that predicts the probability of an event occurring using explanatory variables (each data item), and can be expressed by the following equation (1). log( P / (1-P) ) = B_0 + B_1X_1 + B_2X_2 + ... ···(1) Here, P is the probability of employee turnover, X_1, X_2,... are each data item (explanatory variable), B_0 is the intercept (constant term), and B_1, B_2,... are coefficients (partial regression coefficients) that indicate the magnitude of the influence of each data item on the probability of employee turnover. The model is trained using past employee turnover data and optimizes this coefficient B. Then, by substituting the data of the employee to be analyzed into this formula, the employee's probability of turnover P is calculated. By looking at the calculated coefficient values, it is possible to visualize which items have a positive (plus) or negative (minus) influence on turnover, and to what extent, which leads to the identification of turnover risk factors. Note that the model used is not limited to this, and other classification algorithms (e.g., decision trees, support vector machines, neural networks, etc.) may also be used.

[0065] The action recommendation module 2033 instructs the generation AI system 30 to generate recommended actions that the administrator should take, based on the identified risk factors. At this time, the action recommendation module 2033 obtains personality trait data, etc., from the employee data table 2021 and risk factors from the analysis results table 2022, and transmits this information to the generation AI system 30. The generation AI system 30, upon receiving the instructions, uses this information to generate recommended actions and returns the results to the server 20. The action recommendation module 2033 stores the recommended actions received from the generation AI system 30 in the analysis results table 2022. This makes it possible to present the generated recommended actions, along with other analysis results (risks, factors, etc.), to the administrator's terminal device 10.

[0066] The effectiveness measurement module 2034 controls a feedback loop to continuously improve the accuracy of the system. It receives the results of actions entered by the administrator from terminal device 10, analyzes the changes in employee data before and after the action, and measures the effectiveness of the action. Furthermore, it retrains the machine learning model based on the results of this effectiveness measurement.

[0067] <5. Data Structure> The main data structures used in the retirement prediction system of this embodiment will be described with reference to Figures 4 and 5. Note that the table configuration and item names described herein are examples and can be divided, merged, or modified without departing from the spirit of the invention.

[0068] Figure 4 shows an example of the data structure of an employee data table 2021 that can be stored in the memory unit 202 of the server 20. This table aggregates and stores information about employees that will be used as the basis for analysis. Specifically, it includes items such as "Employee ID," "Quantitative Data," "Qualitative Text Data," "Score Data," "Personality Trait Data," and "Growth Image." "Employee ID" is an identifier that uniquely identifies an employee. "Quantitative Data" stores structured numerical data such as survey response history or personnel information. This item is retained not only as data at a single point in time, but also as a time-series history. This makes it possible to capture temporal changes (differences), such as a sudden drop in satisfaction scores, and improves the accuracy of turnover risk analysis. "Personnel Information" refers to data about employee attributes and careers, such as gender, age, occupation, position, length of service, department, employment type, and evaluation history. "Qualitative Text Data" is an optional item that is stored as needed, and stores text data such as dialogue logs obtained from the conversational AI system 40. The "Score Data" section stores quantitative score data converted from "Qualitative Text Data" by the interactive AI system 40. The "Personality Trait Data" section contains optional items that can be stored as needed. For example, it may store multiple parameter values ​​(quantitative data) obtained based on specific methods such as the FFS (Five Factors & Stress) diagnosis, or text descriptions interpreting those values ​​(qualitative data). The FFS diagnosis is a theory and diagnostic method that analyzes an individual's thinking and behavioral characteristics or stress response based on "five factors" (cohesiveness, receptiveness, discriminativeness, diffuseness, and preservation). The "Growth Image" section contains optional items that can be stored as needed. This data indicates what kind of outlook an employee has on their career, and is obtained, for example, from statements made during interviews.

[0069] Figure 5 shows an example of the data structure of the Analysis Results Table 2022. This table records and manages the results generated after AI analysis. Specifically, it includes items such as "Employee ID," "Resignation Risk," "Resignation Risk Factors," "Recommended Action," and "Action Result." "Employee ID" uniquely identifies the employee being analyzed. "Resignation Risk" stores a score indicating the calculated probability of the employee resigning. "Resignation Risk Factors" stores information indicating the specific factors that formed the basis of the calculated resignation risk.

[0070] Note that "Recommended Actions" and "Action Results" are optional items that are stored as needed. "Recommended Actions" stores specific actions that supervisors should take, generated based on identified risk factors, particularly the content of interview support information. This interview support information includes specific interview procedures (icebreakers, in-depth discussion of issues, consideration of solutions, closing methods, etc.) or recommended questions tailored to the employee's characteristics, taking into account the employee's personality traits data (e.g., FFS assessment results). In addition, actions deemed highly effective based on data of past actions and their effects may also be recommended.

[0071] The "Action Results" section is for recording and accumulating the results after managers have implemented recommended actions (such as interviews). This may include structured evaluations such as changes in the employee's behavior after the interview (e.g., options like "became more positive" or "no particular change"), or free-form summaries of the interview written by the manager. The information recorded here (especially success stories) will be used for subsequent effectiveness measurement or for retraining machine learning models.

[0072] <6. Operation> The operation of the retirement prediction system 1 in this embodiment will now be described. Figure 6 is a flowchart showing an example of the basic flow in which each module of the server 20 in this embodiment works together to perform processing. The control unit 203 of the server 20 performs functions such as the data acquisition module 2031 and the risk prediction module 2032.

[0073] In step S11, the server 20 receives an instruction from the administrator to operate the terminal device 10 to start an analysis on a specific employee (received instruction to start analysis).

[0074] In step S12, the data acquisition module 2031 acquires quantitative data of the employee to be analyzed and quantitative score data already scored by the conversational AI system 40 from an external system and stores them in the employee data table 2021 (data acquisition).

[0075] In step S13, the risk prediction module 2032 instructs the generating AI system 30 to perform an analysis using the data acquired in step S12, and receives the employee turnover risk as a result. The risk prediction module 2032 can also optionally identify the underlying turnover risk factors. The calculated results are stored in the analysis results table 2022 (risk calculation and factor identification).

[0076] In step S14, the control unit 203 of the server 20 formats the analysis results (retirement risk, risk factors, etc.) stored in the analysis results table 2022 into a format for display on the terminal device 10, and transmits it to the terminal device 10 via the communication unit 201. The administrator confirms these results on the screen of the terminal device 10 (results presentation).

[0077] In addition to the basic processing flow described above, the system of this embodiment can be equipped with optional extensions to support the utilization of analysis results and to continuously improve the accuracy of the system, as described below.

[0078] For example, the control unit 203 can use the action recommendation module 2033 to search for and utilize data on past actions and their effects based on the resignation risk factors identified in step S14. For instance, if there is data showing that actions such as "assignment changes" or "career consultations" were previously performed on other employees with similar risk factors and resulted in positive changes, the control unit 203 can recommend similar actions to the employee currently being analyzed based on that track record. This supports data-driven and effective interventions that do not rely solely on intuition or experience.

[0079] <7. Screen example> Figure 7 shows an example of the "Resignation Risk Analysis Screen" displayed on the terminal device 10's display 141 when a manager checks the resignation risk of a specific subordinate employee. This screen aggregates the analysis results performed by the server 20 and is designed to allow managers to intuitively and comprehensively understand the employee's situation.

[0080] At the top of the screen, basic information such as the name or department of the subordinate employee being analyzed is displayed. Below that, the main analysis results are displayed, divided into several areas.

[0081] Area 1411 is the area that displays the "probability of employee turnover." Here, the probability that the employee will leave within a specified period, calculated by the employee turnover prediction model, is displayed in large numbers, such as "92%." This quantifies how well the employee's data matches the trends of past employees who have left the company, allowing for a quick understanding of the severity of the risk.

[0082] Area 1412 is the area that displays the "Alert Distribution." Here, items in which employees have exceeded the risk threshold are aggregated and displayed by data category, such as "Dialogue Log Analysis," "ES (Employee Satisfaction Survey)," "Evaluation FB (Evaluation Feedback Survey)," "Mental Health Survey," and "Intention to Leave." Particularly important indicators are pre-weighted, allowing managers to understand which areas are concentrated with problems and use this as a reference to determine the priority of responses. In the example in Figure 7, alerts are shown for the items "Dialogue Log Analysis," "ES (Employee Satisfaction Survey)," and "Intention to Leave."

[0083] Area 1413 is the area that displays "turnover risk factors." The basis for the high turnover probability shown in Area 1411, that is, the factors that contributed most to the prediction results, are displayed in a list format. Examples of factors displayed include specific survey and analysis items such as "eNPS," "career: fulfillment: satisfaction," and "sense of growth: job content suitability." This allows managers to understand the basis for the prediction of "why the risk is high" and to consider countermeasures with a sense of conviction.

[0084] <8.Summary> As described above, according to the employee turnover prediction system of this embodiment, the server 20 first acquires "quantitative data" such as employee surveys. In addition, it acquires "quantitative score data" converted from "qualitative data" such as dialogue logs that contain employees' true feelings. Then, using this quantitative data and score data as input, the server calculates the employee turnover risk using an "employee turnover prediction model" built based on past data. This employee turnover prediction model can be selected according to the purpose, such as a machine learning model with high accuracy or a statistical analysis model with excellent interpretability. In this way, by combining objective indicators with information that reflects the internal state of employees, it becomes possible to capture signs of turnover with higher accuracy than conventional methods.

[0085] Furthermore, in this embodiment, the server 20 not only indicates the level of risk but also specifically presents the "resignation risk factors" that underlie it, thereby preventing the prediction results from becoming a black box and providing managers with convincing insights. In addition, based on the identified risk factors, employee personality traits, and past success case data, it generates and presents "recommended actions" optimized for each individual (e.g., specific interview procedures and question items). As a result, managers can plan and execute effective resignation prevention actions with clear, data-driven evidence, contributing to an improvement in the overall employee retention rate of the organization.

[0086] (modified version) The above embodiment described a basic configuration for calculating employee turnover risk and presenting its contributing factors. This modified version describes extended functionality that, based on the analysis results, more specifically supports managers' turnover prevention actions and further measures the effectiveness of those actions to continuously improve the overall accuracy of the system.

[0087] <Generate recommended actions> In this modified example, the action recommendation module 2033 in the control unit 203 of the server 20 instructs the generating AI system 30 to generate specific "recommended actions" that the administrator should take, according to the identified risk factors. These "recommended actions" are a concept that encompasses "interview support information" to assist in dialogue with employees and "proposals for specific personnel measures" which are more direct intervention measures.

[0088] Action Recommendation Module 2033 first retrieves the "resignation risk factors" of the target employee from Analysis Results Table 2022. For example, if the risk factor is related to interpersonal relationships such as "communication with supervisor," then "interview support information" is primarily generated. On the other hand, if the risk factor is related to more structural problems such as "mismatch in job content" or "lack of growth opportunities," then "proposals for specific HR measures" may also be generated.

[0089] When generating "interview support information," the action recommendation module 2033 retrieves the employee's "personality trait data" and "growth image" from the employee data table 2021, integrates these with risk factors, and constructs instructions (prompts) for the generating AI system 30. Upon receiving these instructions, the generating AI system 30 uses its advanced language generation capabilities to generate detailed interview procedures optimized for each target employee, as well as specific questions that take into account the employee's characteristics. This enables managers to conduct high-quality interviews with a deep understanding of their subordinates' situations in advance.

[0090] When generating "proposals for specific HR measures," the action recommendation module 2033 instructs the generation AI system 30 to generate HR intervention measures that go beyond interviews, depending on the content of the risk factors. For example, if the risk factor is "mismatch in job content," it will generate "consideration of assignment change" as a recommended action, and if the risk factor is "lack of growth opportunities," it will generate "career support (such as proposing training or setting up a mentor)" as a recommended action.

[0091] The multifaceted recommended actions generated in this way are stored in the analysis results table 2022, then transmitted to the terminal device 10, and presented to the administrator. This allows the administrator to plan and execute consistent employee retention activities, from building relationships through dialogue to specific environmental improvements.

[0092] Figure 8 shows an example of "interview support information" that may be displayed on the terminal device 10's display 141 when a manager checks the resignation risk of a specific subordinate employee and conducts an interview. This screen aggregates the analysis results performed by the server 20 and is designed to enable managers to prevent employee resignation through high-quality communication.

[0093] The screen is mainly divided into three areas: "Individual Basic Information," "Summary of Problem Factors," and "How to Conduct the Interview."

[0094] Area 1414 of the "Individual Basic Information" section displays the personality traits data of the employee being analyzed, as well as their desired "growth image." Here, "personality traits data" refers to the characteristics of the employee's thinking and behavior, based on, for example, the results of an FFS (Five Factors & Stress) assessment. However, "personality traits data" is not limited to FFS assessment results. In the example in Figure 8, the classification is displayed as "CEB," which indicates that the individual possesses a combination of C trait (discriminativeness: a logical and efficiency-oriented trait), E trait (preservativeness: a trait that aims for steady growth and improvement), and B trait (receptiveness: a trait that is highly empathetic and sensitive to changes in the surroundings). Furthermore, the "growth image" indicates the employee's outlook on their career. For example, if it is displayed as "gradual," it indicates a tendency to prefer steady growth over rapid change. This allows managers to gain a deeper understanding of the employee's personality and career aspirations before interviews.

[0095] In area 1415, "Summary of Problem Factors," a summary of employee problems, analyzed from qualitative data by the conversational generation AI system 40, is displayed. In the example in Figure 8, the issues are organized into categories such as "Problems with Job Satisfaction" and "Problems with Interpersonal Relationships." Each category displays specific problem items, such as "I don't feel fulfilled at work," along with a score indicating its severity and the number of items (e.g., "1 / 6"). This allows managers to identify issues that should be focused on during interviews, both in advance and quantitatively.

[0096] In section 1416, "How to Conduct an Interview," specific steps are presented chronologically, starting with an icebreaker and then delving deeper into each source of concern. Each step includes specific recommended questions, but these are not merely template responses. Action Recommendation Module 2033 dynamically generates the wording and content of questions, taking into account the employee's personality trait data (e.g., FFS assessment results) shown in the "Individual Basic Information" section, and the employee's desired "growth image." For example, for an employee with a trait that emphasizes logic (C trait), questions focusing on specific facts and efficiency are presented to make it easier to elicit honest opinions. In this way, by supporting personalized communication tailored to each employee, managers can conduct more effective interviews. In the final step, "Considering Solutions," managers are encouraged to work with the employee to develop specific action plans for the issues revealed in the interview, such as "areas that can be improved in the current situation" or "how the company can provide support."

[0097] <Effectiveness measurement and relearning> In this modified example, after the administrator implements a recommended action such as an interview or assignment change, they input the results or employee reactions from the terminal device 10. This action result data is received by the effectiveness measurement module 2034 of the server 20, and feedback processing is performed. The action result data entered by the administrator here includes specific details of the action taken, such as "what kind of words or suggestions were made" or "whether an assignment change was implemented," as well as an evaluation of the change in the employee's state as a result, such as "the employee's attitude became more positive" or "there was no particular change (the negative situation continued)." Alternatively, a questionnaire may be administered to the employee after the interview with their supervisor, and the results of the questionnaire may be obtained as action result data.

[0098] The effectiveness measurement module 2034 measures the effectiveness of a recommended action by comparing quantitative and score data before and after the action is implemented. For example, consider a scenario where "lack of growth opportunities" is identified as a risk factor for an employee's resignation, and the supervisor conducts a career interview (recommended action). The effectiveness measurement module 2034 first compares whether the "perceived growth" item (quantitative data) in a survey, which was "2" (on a 5-point scale) before the action, improved to "4" three months after the action was implemented. Furthermore, this system can also analyze text data obtained from questionnaires, such as the free-response section of a survey, using the generating AI. It also compares changes such as whether the "career concern score" (score data) generated from free-response comments (qualitative text data) such as "I am worried about my future career" in the pre-action survey was "5" (high concern), while in the post-action survey, the score decreased to "1" (low concern) based on a comment such as "I saw a direction in the interview." In this way, by capturing changes before and after using both structured quantitative data and score data generated from text, it becomes possible to objectively evaluate whether the actions taken have brought about positive changes in employees' psychological states and to measure their effects.

[0099] Next, based on the results of the effectiveness measurement, particularly successful cases where "positive changes were observed" (i.e., data combining specific actions and positive results), these are used as new training data to further train (retrain) the machine learning model. Through this series of processes, the system accumulates insights into which actions were effective for which types of employees, continuously improving not only the accuracy of predicting employee turnover risk but also the accuracy of recommending appropriate actions.

[0100] <Risk calculation process by the server> In the above embodiment, a configuration was described in which an external generation AI system 30 executes a retirement prediction model based on instructions from the server 20, but the present invention is not limited thereto. For example, the server 20 itself may hold the retirement prediction model, and the processor 29 of the server 20 may execute a process to calculate retirement risk.

[0101] In this configuration, the risk prediction module 2032, located in the control unit 203 of the server 20, executes the employee turnover prediction model as its own function, rather than outsourcing the processing to the external generation AI system 30. Specifically, the risk prediction module 2032 directly inputs the "quantitative data" and "score data" collected by the data acquisition module 2031, and calculates the employee turnover risk using the employee turnover prediction model (e.g., a statistical analysis model or a machine learning model) stored in the memory unit 202. The calculated employee turnover risk and the underlying risk factors are stored in the analysis results table 2022 and then output to the terminal device 10 used by the administrator.

[0102] <Basic Computer Hardware Configuration> Figure 9 is a block diagram showing the basic hardware configuration of computer 90. Computer 90 includes at least a processor 901, main memory 902, auxiliary storage 903, and a communication interface IF991. These are electrically connected to each other by a communication bus.

[0103] The processor 901 is hardware for executing the instruction set written in a program. The processor 901 consists of an arithmetic unit, registers, peripheral circuits, etc.

[0104] Main memory 902 is used to temporarily store programs and data processed by programs, etc. For example, it is a volatile memory such as DRAM (Dynamic Random Access Memory).

[0105] Auxiliary storage device 903 refers to a storage device for saving data and programs. Examples include flash memory, HDD (Hard Disc Drive), magneto-optical disk, CD-ROM, DVD-ROM, and semiconductor memory.

[0106] The IF991 communication interface is an interface for inputting and outputting signals for communication with other computers via a network using wired or wireless communication standards. A network consists of various mobile communication systems, such as the internet, LANs, and wireless base stations. For example, a network includes 3G, 4G, and 5G mobile communication systems, LTE (Long Term Evolution), and wireless networks that can connect to the internet via designated access points (e.g., Wi-Fi®). When connecting wirelessly, communication protocols include, for example, Z-Wave®, ZigBee®, and Bluetooth®. When connecting via a wired connection, the network also includes connections made directly via USB (Universal Serial Bus) cables, etc.

[0107] Furthermore, by distributing all or part of each hardware configuration across multiple computers 90 and connecting them to each other via a network, a computer 90 can be virtually realized. Thus, the concept of computer 90 includes not only a computer 90 housed in a single enclosure or case, but also a virtualized computer system.

[0108] <Basic Functional Configuration of Computer 90> The functional configuration of the computer realized by the basic hardware configuration of computer 90 (Figure 9) will be explained. The computer comprises at least one functional unit: a control unit, a memory unit, and a communication unit.

[0109] Furthermore, the functional units of computer 90 can also be realized by distributing all or part of each functional unit across multiple computers 90 interconnected via a network. The concept of computer 90 includes not only a single computer 90 but also a virtualized computer system.

[0110] The control unit is realized when the processor 901 reads various programs stored in the auxiliary storage device 903, loads them into the main memory device 902, and executes processing according to those programs. The control unit can realize various functional units that perform information processing depending on the type of program. In this way, the computer is realized as an information processing device that performs information processing.

[0111] The memory unit is implemented by the main memory 902 and the auxiliary memory 903. The memory unit stores data, various programs, and various databases. The processor 901 can also reserve memory areas corresponding to the memory unit in the main memory 902 or the auxiliary memory 903 according to the program. The control unit can also cause the processor 901 to perform operations such as adding, updating, and deleting data stored in the memory unit according to the various programs.

[0112] A database, specifically a relational database, is used to manage and link together tabular data sets called masters, which are structurally defined by rows and columns. In a database, tables are called tables, masters are called masters, the columns of tables are called columns, and the rows of tables are called records. In a relational database, relationships can be established and linked between tables and masters. Typically, each table and master has a primary key column to uniquely identify records, but setting a primary key column is not mandatory. The control unit can instruct the processor 901 to add, delete, or update records in specific tables and masters stored in the memory unit, according to various programs. Furthermore, by storing data, various programs, and various databases in the memory unit, the information processing device and information processing system related to this disclosure can be considered to have been manufactured.

[0113] Furthermore, the databases and masters in this disclosure may include any data structures (lists, dictionaries, associative arrays, objects, etc.) in which information is structurally defined. Data structures also include data that can be considered as data structures by combining data with functions, classes, methods, etc. written in any programming language.

[0114] The communication unit is implemented by the communication IF991. The communication unit provides the functionality to communicate with other computers 90 via the network. The communication unit can receive information transmitted from other computers 90 and input it to the control unit. The control unit can cause the processor 901 to perform information processing on the received information according to various programs. The communication unit can also transmit information output from the control unit to other computers 90.

[0115] Furthermore, each of the above-mentioned configurations, functions, processing units, processing means, etc., may be implemented in hardware, in whole or in part, for example, by designing them as integrated circuits. The present invention can also be implemented by software program code that realizes the functions of the embodiment. In this case, a storage medium on which the program code is recorded is provided to a computer, and the processor of that computer reads the program code stored in the storage medium. In this case, the program code read from the storage medium itself realizes the functions of the embodiment described above, and the program code itself and the storage medium on which it is stored constitute the present invention. Examples of storage media used to supply such program code include flexible disks, CD-ROMs, DVD-ROMs, hard disks, SSDs, optical disks, magneto-optical disks, CD-Rs, magnetic tapes, non-volatile memory cards, ROMs, etc.

[0116] Furthermore, the program code that implements the functions described in this embodiment can be implemented in a wide range of programming or scripting languages, such as assembler, C / C++, Perl, Shell, PHP, and Java (registered trademark).

[0117] Furthermore, the program code for the software that implements the functions of the embodiment may be distributed via a network and stored in a storage means such as a computer's hard disk or memory, or in a storage medium such as a CD-RW or CD-R, and the computer's processor may read and execute the program code stored in the storage means or storage medium.

[0118] The functions realized by the components described herein may be implemented in a circuit or processing circuitry, including general-purpose processors, application-specific processors, integrated circuits, ASICs (Application Specific Integrated Circuits), CPUs (a Central Processing Unit), conventional circuits, and / or combinations thereof, programmed to realize the functions described herein. A processor is considered a circuit or processing circuitry, including transistors and other circuits. A processor may be a programmed processor that executes a program stored in memory. In this specification, circuitry, unit, and means are hardware programmed to perform or execute the functions described herein. Such hardware may be any hardware disclosed herein, or any hardware known to be programmed to perform or execute the functions described herein. If the hardware is a processor that is considered to be a type of circuitry, then the circuitry, means, or unit is a combination of hardware and software used to constitute the hardware and / or processor.

[0119] While several embodiments of this disclosure have been described above, these embodiments can be implemented in a variety of other forms, and various omissions, substitutions, and modifications are permitted without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims and their equivalents.

[0120] (Note) The details described in each of the above embodiments are noted below.

[0121] (Note 1) A retirement prediction program for operating a computer comprising a processor and memory, the program causing the processor to perform the following steps: acquiring quantitative data on an employee's work attitudes; acquiring quantitative score data converted from qualitative data on an employee's work attitudes using a generating AI; taking the quantitative data and score data as input and using a retirement prediction model constructed based on past retirement data to calculate the risk of an employee resigning within a predetermined period; and outputting the retirement risk to a terminal device used by an administrator. (Note 2) The program described in (Appendix 1) outputs the retirement risk in a step that associates the retirement risk with the underlying retirement risk factors. (Note 3) The program described in (Appendix 2) calculates employee turnover risk by assigning different weights to specific employee turnover risk factors that have been predetermined as key indicators, compared to other employee turnover risk factors, in the step of calculating employee turnover risk. (Note 4) The employee turnover prediction model is a program described in (Appendix 2) or (Appendix 3), which uses statistical testing to identify items that have a significant correlation with the employee turnover rate as employee turnover risk factors, and calculates the employee turnover risk based on whether or not the identified items exceed a predetermined threshold. (Note 5) The employee turnover prediction model is a machine learning model trained on quantitative and score data of past employee turnovers, and is one of the programs described in any one of the items (Appendix 2) to (Appendix 4). (Note 6) A program described in any one of the items (Appendix 1) to (Appendix 5), which calculates retirement risk based on the change in quantitative data or score data from a predetermined point in the past to the present in the step of calculating retirement risk. (Note 7) A program as described in any one of the items (Appendix 2) to (Appendix 6), which causes the processor to perform the step of generating and outputting recommended actions that the manager should take based on identified retirement risk factors. (Note 8) The program described in (Appendix 7) includes a step of generating recommended actions, which involves generating interview support information, including recommended questions to be asked during interviews with employees, based on identified employee turnover risk factors and pre-stored employee personality trait data. (Note 9) The program described in (Appendix 7) or (Appendix 8) generates recommended actions that the administrator should take, based on data regarding previously performed recommended actions and their effects, in the step of generating recommended actions. (Note 10) A program as described in any one of the items (Appendix 7) to (Appendix 9), which causes the processor to perform the step of measuring the change in the employee's psychological state after the recommended action has been taken as an effect, and storing the recommended action in association with the effect. (Note 11) The program described in (Appendix 10) further causes the processor to perform the step of retraining a model for generating retirement prediction models or recommended actions using the effects. (Note 12) An information processing device that executes all steps in the program described in any one of the items (Appendix 1) to (Appendix 11). (Note 13) A method for performing all steps in the retirement prediction program described in any one of the items (Appendix 1) to (Appendix 11). (Note 14) A system comprising means for executing all steps in any one of the items (Appendix 1) to (Appendix 11). [Explanation of Symbols]

[0122] 1... System 10…Terminal device 12…Communication IF 13…Input device 14…Output device 15…Memory 16…Storage 19… Processor 20... Server 22...Communication IF 23…Input / Output Interface 25…Memory 2 hours… storage 29… Processor 30…Generating AI system 40…Conversational AI system 50…Talent Management System 80…Network

Claims

1. A retirement prediction program for operating a computer comprising a processor and memory, The aforementioned processor, Steps to obtain quantitative data on employees' work attitudes, The process involves obtaining quantitative score data converted using generative AI from qualitative data regarding employees' work attitudes, and The steps include: using the aforementioned quantitative data and the aforementioned score data as input, and using a retirement prediction model constructed based on past retirement data, to calculate the retirement risk of an employee leaving the company within a predetermined period; The steps include identifying the retirement risk factors that form the basis of the aforementioned retirement risk, A step of generating the content of recommended actions that the manager should take based on the aforementioned employee turnover risk factors and the employee's personality trait data that has been stored in advance, A program that performs the steps of outputting the aforementioned employee turnover risk, employee turnover risk factors, and recommended actions to a terminal device used by the administrator.

2. The program according to claim 1, wherein in the step of calculating the retirement risk, a specific retirement risk factor predetermined as an important indicator is weighted differently from other retirement risk factors, and the retirement risk is calculated.

3. The program according to claim 1, wherein the retirement prediction model is a model that uses statistical testing to identify items that have a significant correlation with the retirement rate as retirement risk factors, and calculates the retirement risk based on whether or not the identified items exceed a predetermined threshold.

4. The program according to claim 1, wherein the retirement prediction model is a machine learning model that has learned the quantitative data and score data of past retirees.

5. The program according to claim 1, wherein in the step of calculating the retirement risk, the retirement risk is calculated based on the amount of change in the quantitative data or score data from a predetermined point in the past to the present.

6. The program according to claim 1, wherein the step of generating the recommended action includes generating interview support information, which includes recommended questions to be asked during an interview with the employee, based on the identified employee resignation risk factors and pre-stored employee personality trait data.

7. The program according to claim 1, wherein, in the step of generating the recommended action, the program generates a recommended action that the administrator should perform based on data relating to previously performed recommended actions and their effects.

8. The aforementioned processor further includes, The program according to claim 1, which includes the step of measuring the change in the employee's psychological state after the recommended action has been performed as an effect, and storing the recommended action in association with the effect.

9. The aforementioned processor further includes, The program according to claim 8, which performs the step of retraining the retirement prediction model or the model for generating the recommended action using the aforementioned effect.

10. An information processing device that performs all steps in the program described in any one of claims 1 to 9.

11. A method for predicting retirement performed on a computer, comprising performing all steps of a retirement prediction program according to any one of claims 1 to 9.

12. A system comprising means for performing all steps in a program according to any one of claims 1 to 9.

Images