Employee occupational injury risk assessment method based on AI large model

Abstract

The invention relates to the field of administrative management, and provides an employee occupational injury risk assessment method based on an AI large model, and the method comprises the steps: obtaining the group information of a to-be-assessed group, and constructing a first risk vector corresponding to the to-be-assessed group according to the group information; obtaining member information of the to-be-evaluated group, and determining an evaluation scheme of the to-be-evaluated group according to the member information based on a preset rule engine; obtaining an assessment result of the to-be-assessed group based on the assessment scheme, and constructing a second risk vector corresponding to the to-be-assessed group based on the assessment result; and inputting the first risk vector and the second risk vector into a preset risk assessment model to obtain an occupational injury risk corresponding to the to-be-assessed group, and matching an insurance scheme of the to-be-assessed group according to the occupational injury risk.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence, and in particular to a method for assessing employee work injury risks based on a large AI model. Background Technology

[0002] With companies placing greater emphasis on employee benefits and the continuous development of the insurance market, more and more companies are insuring their employees in groups to mitigate potential work-related injury risks such as accidents and occupational diseases. However, large companies have diverse workforces, and different subsidiaries, and even different departments within the same subsidiary, face varying types and levels of risk. Therefore, assessing the work-related injury risks of different employee groups and selecting appropriate insurance products for them has become a pressing issue. Summary of the Invention

[0003] The main purpose of this application is to provide a method for assessing employee work injury risks based on an AI-powered large model, which aims to improve the intelligence of work injury risk management by selecting appropriate insurance products for different employee groups.

[0004] Firstly, this application provides a method for assessing employee work-related injury risks based on an AI-powered large-scale model, the method comprising the following steps: Obtain the group information of the group to be evaluated, and construct a first risk vector corresponding to the group to be evaluated based on the group information; Obtain the member information of the group to be evaluated, and determine the evaluation scheme of the group to be evaluated based on the member information using a preset rule engine; The assessment results of the group to be assessed are obtained based on the assessment scheme, and a second risk vector corresponding to the group to be assessed is constructed based on the assessment results. The first risk vector and the second risk vector are input into a preset risk assessment model to obtain the work injury risk corresponding to the group to be assessed, and an insurance plan for the group to be assessed is matched according to the work injury risk.

[0005] In some implementations, obtaining the group information of the group to be evaluated and constructing a first risk vector corresponding to the group to be evaluated based on the group information includes: Capture work injury case information and perform natural language analysis on the work injury case information, and match each work injury case information with a preset industry type; Based on the work injury case information corresponding to the preset industry types, determine the industry risk vector corresponding to each preset industry type; Determine the target industry type for matching the group information from the preset industry types; Based on the industry risk vector of the target industry type, determine the first risk vector corresponding to the group to be evaluated.

[0006] In some implementations, the member information includes: member department information; before obtaining the member information of the group to be evaluated and determining the evaluation scheme of the group to be evaluated based on the member information using a preset rule engine, the method further includes: Based on the member department information, keyword matching is performed on the member departments of the group to be evaluated to obtain the occupational types corresponding to each member department; Obtain the risk base corresponding to each occupation type. The risk base includes: random risk base and cumulative risk base.

[0007] In some implementations, the member information further includes: member age information; the step of obtaining the member information of the group to be evaluated and determining the evaluation scheme of the group to be evaluated based on the member information using a preset rule engine includes: Based on the member age information and the randomness risk base, determine the randomness risk information corresponding to each group member; Based on the member age information and the cumulative risk base, determine the cumulative risk information corresponding to each group member; The evaluation scheme for each group member is determined based on the randomness risk information and the cumulative risk information.

[0008] In some implementations, determining the assessment scheme corresponding to each group member based on the randomness risk information and the cumulative risk information includes: The random risk information and cumulative risk information of each group member are normalized respectively, and the normalized random risk information is used as the horizontal axis and the normalized cumulative risk information is used as the vertical axis to determine the position of each group member in the preset coordinate system. The evaluation scheme corresponding to the group member is determined based on the position of the group member in the preset coordinate system.

[0009] In some implementations, determining the evaluation scheme corresponding to the group member based on the group member's position in the preset coordinate system includes: The assessment scheme corresponding to the group members in the first quadrant of the preset coordinate system is determined as: first questionnaire assessment and first sampling physical examination; The assessment scheme corresponding to the group members in the second quadrant of the preset coordinate system is determined as: second questionnaire assessment and first sampling physical examination; The assessment scheme corresponding to the group members in the third quadrant of the preset coordinate system is determined as: second questionnaire assessment and second sampling physical examination; The assessment scheme corresponding to the group members in the fourth quadrant of the preset coordinate system is determined as: first questionnaire assessment and second sampling physical examination. The sampling probability of the first physical examination is greater than that of the second physical examination, and the questionnaire complexity of the first questionnaire assessment is greater than that of the second questionnaire assessment.

[0010] In some embodiments, the method further includes: Obtain the questionnaire complexity corresponding to the questionnaire assessment, and determine multiple target assessment rules from each subset of assessment rules in the preset assessment rule set based on the questionnaire complexity. The subset of assessment rules includes: lifestyle habits subset, work habits subset, and mental health subset. Multiple questionnaires are generated based on the target evaluation rules. The questionnaires include: question text and answer options. The questionnaire information generated according to different target evaluation rules is randomly mixed to obtain the survey questionnaire corresponding to the questionnaire evaluation.

[0011] In some implementations, determining the random risk information corresponding to each group member based on the member's age information and the random risk base includes: in, The randomness risk information refers to the maximum age of a member and the minimum age of a member. As the first risk reduction factor, Let x be the randomness risk base, a be the member age information, b be the first preset age, and b be the second preset age.

[0012] In some implementations, determining the cumulative risk information corresponding to each group member based on the member's age information and the cumulative risk base includes: in, The cumulative risk information refers to the maximum age of the member and the minimum age of the member. This is the second risk reduction coefficient. Let x be the cumulative risk base, x be the member's age information, and α be the risk modulation index, where 0 < α < 1.

[0013] In some implementations, the step of inputting the first risk vector and the second risk vector into a preset risk assessment model to obtain the work injury risk corresponding to the group to be assessed includes: The first risk vector and the second risk vector are input into the input layer of the risk assessment model to obtain the encoded input information; The input information is input into at least one hidden layer, and the hidden information is obtained by weighted summation and activation function corresponding to the hidden layer. The hidden information is input into the output layer to obtain the work injury risk level determined by the output layer based on the hidden information; Based on the pre-defined correspondence between work injury risk levels and insurance plans, the corresponding insurance plan for the work injury risk level is determined.

[0014] This application provides a method for assessing employee work injury risks based on an AI-powered large-scale model. The method involves: acquiring group information of the group to be assessed; constructing a first risk vector corresponding to the group based on the group information; acquiring member information of the group to be assessed; determining an assessment scheme for the group based on the member information using a preset rule engine; obtaining the assessment result of the group based on the assessment scheme; constructing a second risk vector corresponding to the group based on the assessment result; inputting the first and second risk vectors into a preset risk assessment model to obtain the work injury risk corresponding to the group to be assessed; and matching an insurance scheme for the group to be assessed based on the work injury risk. By assessing the risks of the group as a whole and the characteristics of each individual member within the group separately, and adjusting the assessment scheme according to the characteristics of the member information, a personalized insurance scheme for different groups to be assessed is determined, thus realizing personalized group insurance schemes. Furthermore, precise insurance reduces work injury risks and improves the intelligence of work injury risk management. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A flowchart illustrating an employee work injury risk assessment method based on an AI large model, provided as an embodiment of this application; Figure 2 This is a schematic block diagram of the structure of a computer device according to an embodiment of this application. Detailed Implementation

[0017] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0018] The flowchart shown in the attached diagram is for illustrative purposes only and does not necessarily include all content and operations / steps, nor does it necessarily have to be performed in the order described. For example, some operations / steps can be broken down, combined, or partially merged, so the actual execution order may change depending on the actual situation.

[0019] This application provides a method for assessing employee work injury risk based on an AI large model.

[0020] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0021] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating an AI-based large-scale model-based employee work injury risk assessment method, provided as an embodiment of this application. This AI-based large-scale model-based employee work injury risk assessment method can be used in terminals or servers to determine the corresponding insurance coverage for different groups to be assessed, thereby achieving personalized group insurance services and reducing work injury risks through precise insurance coverage. The terminal can be an electronic device such as a mobile phone, tablet, laptop, desktop computer, personal digital assistant, or wearable device; the server can be a standalone server, a server cluster, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0022] like Figure 1 As shown, the employee work injury risk assessment method based on AI big model includes steps S101 to S104.

[0023] Step S101: Obtain the group information of the group to be evaluated, and construct the first risk vector corresponding to the group to be evaluated based on the group information.

[0024] For example, group information is basic information that insurance companies need to obtain when contacting a group to be assessed. This includes the organization's full name, address, business scope, nature of the organization (e.g., state-owned, private), industry category, information of the legal representative or person in charge, type and number of the organization's documents (e.g., business license number), validity period of the documents, organization code, and financial information. This information is used to determine the legitimacy of the group and assess its creditworthiness. Through this information, insurance companies can gain a preliminary understanding of the group's risk and determine whether further contact is necessary.

[0025] For example, a first risk vector is determined based on group information to reflect the risk of the group to be insured as a whole.

[0026] In some implementations, obtaining the group information of the group to be evaluated and constructing a first risk vector corresponding to the group to be evaluated based on the group information includes: Capture work injury case information and perform natural language analysis on the work injury case information, and match each work injury case information with a preset industry type; Based on the work injury case information corresponding to the preset industry types, determine the industry risk vector corresponding to each preset industry type; Determine the target industry type for matching the group information from the preset industry types; Based on the industry risk vector of the target industry type, determine the first risk vector corresponding to the group to be evaluated.

[0027] For example, risk levels vary significantly across different industries. Employees in high-risk industries such as construction and mining face higher occupational risks and a greater probability of accidents and work-related injuries; employees in the chemical industry may face a higher risk of occupational diseases; while occupational risks in industries such as finance and the internet are relatively lower. Therefore, based on industry information, insurance companies can classify and assess the risks of insured individuals according to industry characteristics.

[0028] Specifically, relevant reports and information on work-related injury cases are crawled from relevant websites. Natural language processing is used to determine the preset industry types corresponding to the work-related injury case information, and industry risk vectors are obtained for each preset industry type. The industry risk vectors may include, for example, risk probability and risk severity. The preset industry types can be derived from industry directories published by relevant organizations.

[0029] For example, the target industry type can be matched based on group information, which can include the group's department name, job responsibilities, job title, etc. The target industry type for matching the group information can be determined from the preset industry types based on the Big Prophecy model.

[0030] Step S102: Obtain the member information of the group to be evaluated, and determine the evaluation scheme of the group to be evaluated based on the member information using a preset rule engine.

[0031] For example, the member information of the group to be evaluated is the basic information of the group members who need to be insured, such as the insured's name, gender, age, ID number, department, contact information, etc.

[0032] Within a company, individuals are typically divided into departments with different functions, such as administrative, production, and support departments. Assigning group members to specific departments facilitates a more detailed and accurate assessment of the risk for each individual within the insured group.

[0033] In some implementations, the member information includes: member department information; before obtaining the member information of the group to be evaluated and determining the evaluation scheme of the group to be evaluated based on the member information using a preset rule engine, the method further includes: Based on the member department information, keyword matching is performed on the member departments of the group to be evaluated to obtain the occupational types corresponding to each member department; Obtain the risk base corresponding to each occupation type. The risk base includes: random risk base and cumulative risk base.

[0034] For example, a company's internal department division is usually based on functions. However, different companies' department divisions are determined by the company's development needs, which is different from the division method required for life insurance risk assessment. Different companies have different basis for department division. In order for member department information to be used for life insurance risk assessment, it is necessary to determine the occupational type corresponding to each member department based on the member department information.

[0035] Specifically, a list of common occupational names or job titles can be pre-defined. The information of each member department can then be matched with the contents of this list to determine the corresponding occupational information for each department. For example, the occupational information for Mining Department 1 and Mining Department 2 is mining operators; the occupational information for Human Resources Department and Finance Department is administrative staff; and the occupational information for Canteen Management Department and Dormitory Management Department is logistics staff.

[0036] For example, the risk base is used to reflect the risk level corresponding to different occupational types. A risk base is pre-set for each occupational information. The risk base can be obtained from existing research using a large language model, or it can be obtained from a survey commissioned to a third-party organization; no limitation is made here. The risk base includes: a random risk base and a cumulative risk base. The random risk base reflects the probability of risk occurrence due to unexpected situations, while the cumulative risk base reflects the probability of risk occurrence due to pathogenic factors. For instance, occupations such as drivers and high-altitude workers have a certain probability of accidents, resulting in a relatively high random risk base. However, these occupations have fewer pathogenic factors with cumulative effects, resulting in a relatively low cumulative risk base. Occupations such as chemical researchers have a certain probability of exposure to toxic substances, therefore, these occupations have a relatively high cumulative risk base, but due to the low level of danger, their random risk base is relatively low. Occupations such as miners are exposed to dusty environments, and accidents may occur in the working environment; therefore, these occupations have both a high random risk base and a high cumulative risk base.

[0037] In some implementations, the member information further includes: member age information; the step of obtaining the member information of the group to be evaluated and determining the evaluation scheme of the group to be evaluated based on the member information using a preset rule engine includes: Based on the member age information and the randomness risk base, determine the randomness risk information corresponding to each group member; Based on the member age information and the cumulative risk base, determine the cumulative risk information corresponding to each group member; The evaluation scheme for each group member is determined based on the randomness risk information and the cumulative risk information.

[0038] For example, the probability of different types of risks occurring is related to age, and the relationship between different types of risks and members' ages varies. For instance, as age increases, operational proficiency and mental capacity improve, leading to a decrease in the probability of random risks occurring; however, for cumulative risks, the probability increases with age-related decline in physical skills and the accumulation of pathogenic factors. Therefore, random risk information for group members is determined based on member age information and the random risk baseline, reflecting the likelihood of a specific group member experiencing a random risk; similarly, cumulative risk information is determined based on member age information and the cumulative risk baseline, reflecting the likelihood of a specific group member experiencing a cumulative risk.

[0039] For example, group members are divided into different types based on random risk information and cumulative risk information, and different assessment schemes are used to assess different types of group members.

[0040] Specifically, the relationships between random risk information and member age information and random risk base, as well as the relationships between cumulative risk information and member age information and cumulative risk base, can all be obtained through fitting.

[0041] In some implementations, determining the assessment scheme corresponding to each group member based on the randomness risk information and the cumulative risk information includes: The random risk information and cumulative risk information of each group member are normalized respectively, and the normalized random risk information is used as the horizontal axis and the normalized cumulative risk information is used as the vertical axis to determine the position of each group member in the preset coordinate system. The evaluation scheme corresponding to the group member is determined based on the position of the group member in the preset coordinate system.

[0042] For example, in order to unify the scale of multiple random risk information and cumulative risk information, the random risk information and cumulative risk information can be normalized separately to the range of [-1, 1], so as to more accurately classify the types of group members according to their positions in the preset coordinate system.

[0043] In some implementations, determining the evaluation scheme corresponding to the group member based on the group member's position in the preset coordinate system includes: The assessment scheme corresponding to the group members in the first quadrant of the preset coordinate system is determined as: first questionnaire assessment and first sampling physical examination; The assessment scheme corresponding to the group members in the second quadrant of the preset coordinate system is determined as: second questionnaire assessment and first sampling physical examination; The assessment scheme corresponding to the group members in the third quadrant of the preset coordinate system is determined as: second questionnaire assessment and second sampling physical examination; The assessment scheme corresponding to the group members in the fourth quadrant of the preset coordinate system is determined as: first questionnaire assessment and second sampling physical examination. The sampling probability of the first physical examination is greater than that of the second physical examination, and the questionnaire complexity of the first questionnaire assessment is greater than that of the second questionnaire assessment.

[0044] Understandably, random risk can be assessed through questionnaires, which score group members’ mental health, work habits, and lifestyle habits to determine the level of random risk; cumulative risk can be assessed through physical examinations, which determine the level of cumulative risk based on whether group members have lesions, sub-health conditions, etc.

[0045] For example, group members in the first quadrant have higher random and cumulative risks, so they need to be underwritten using a sampling medical examination and questionnaire approach, with both the sampling probability of the medical examination and the complexity of the questionnaire being high. Group members in the second quadrant also have high random risks and need to be sampled for medical examination with a high sampling probability, but the complexity of the questionnaire can be simplified. Group members in the third quadrant have lower random and cumulative risks, and the sampling probability of the medical examination and the complexity of the questionnaire can be reduced. Group members in the fourth quadrant have lower cumulative risks and can be sampled for medical examination with a lower sampling probability, but the random risk is high, requiring a high complexity of the questionnaire.

[0046] For example, different sampling probabilities are used to conduct physical examinations on members of different types of groups, avoiding the waste of medical resources and time costs. Similarly, different questionnaire complexities are used for different types of group members, avoiding the waste of time costs for group members and the computational costs required to analyze the questionnaires.

[0047] In some implementations, determining the random risk information corresponding to each group member based on the member's age information and the random risk base includes: in, The randomness risk information refers to the maximum age of a member and the minimum age of a member. As the first risk reduction factor, Let x be the randomness risk base, a be the member age information, b be the first preset age, and b be the second preset age.

[0048] For example, within a certain age range x∈[min, a], the randomness risk decreases with age, where min can be the actual minimum age in the member's age information or the theoretical minimum age allowed to participate in labor, and the value of a can be set according to actual needs, for example, a=min+5. The randomness risk decreases to a certain level. Then, a plateau phase occurs when x∈(a, b), where The value of can be, for example, 0.9, where the value of b can be set according to actual needs, for example, b = max - 10. After the age increases to a certain level, the randomness risk will increase again within the age range x ∈ [b, max], where max can be the actual maximum age in the member's age information, or the theoretical retirement age.

[0049] In some implementations, determining the cumulative risk information corresponding to each group member based on the member's age information and the cumulative risk base includes: in, The cumulative risk information refers to the maximum age of the member and the minimum age of the member. This is the second risk reduction coefficient. Let x be the cumulative risk base, x be the member's age information, and α be the risk modulation index, where 0 < α < 1.

[0050] For example, the probability of cumulative risk gradually increases with age, then tends to a stable value, and the probability of cumulative risk is lower when the age information of members is relatively small; for example, when the age of members is min, the cumulative risk is [value missing]. ,in =0.3; the probability of cumulative risk is higher when the age of members is large. For example, when the age of members is max, the cumulative risk is 0.3. The risk modulation index is used to describe the convergence of cumulative risk information. The degree of smoothness is α=0.5 in the employee work injury risk assessment method based on AI large model provided in this embodiment.

[0051] In some embodiments, the method further includes: Obtain the questionnaire complexity corresponding to the questionnaire assessment, and determine multiple target assessment rules from each subset of assessment rules in the preset assessment rule set based on the questionnaire complexity. The subset of assessment rules includes: lifestyle habits subset, work habits subset, and mental health subset. Multiple questionnaires are generated based on the target evaluation rules. The questionnaires include: question text and answer options. The questionnaire information generated according to different target evaluation rules is randomly mixed to obtain the survey questionnaire corresponding to the questionnaire evaluation.

[0052] For example, the preset set of evaluation rules contains a large number of evaluation rules. These rules reflect the factors that influence the evaluation results. Evaluation rules can be general rules such as sleep habits and dietary habits. To prevent users from giving false or arbitrary answers, multiple questionnaires are generated for each evaluation rule based on a large language model. Only when the answers in the multiple questionnaires do not show significant contradictions is the user's answer accepted. For instance, for the evaluation rule of "sleep habits," multiple questionnaires can be generated, with question texts such as: "Your sleep duration is:", "Your weekday sleep duration is:", "Your rest day sleep duration is:", "Your nighttime sleep duration and afternoon nap duration are:", etc. The multiple questionnaires generated by different evaluation rules are then mixed together.

[0053] Understandably, the number of goal assessment rules is positively correlated with questionnaire complexity, and the number of questionnaire messages generated by the same goal assessment rule is also positively correlated with questionnaire complexity. Specifically, the number of questionnaire messages generated by the same goal assessment rule can be any integer in [1, 10], and the selection of integers follows a normal distribution.

[0054] Step S103: Obtain the evaluation results of the group to be evaluated based on the evaluation scheme, and construct the second risk vector corresponding to the group to be evaluated based on the evaluation results.

[0055] For example, the responses of group members can be scored based on the results of questionnaire assessments, such as scores for lifestyle habits, work habits, and mental health; the physical examination results of group members can also be scored based on sampled physical examinations.

[0056] For example, the physical examination scores and questionnaire scores of group members can be concatenated to obtain the second feature vector corresponding to the second risk factor. Of course, it is not limited to this. The second feature vector may also include the physical examination pass rate, or it may include the numerical indicators of specific physical examination items.

[0057] Step S104: Input the first risk vector and the second risk vector into a preset risk assessment model to obtain the work injury risk corresponding to the group to be assessed, and match the insurance plan for the group to be assessed according to the work injury risk.

[0058] For example, the risk assessment model can be trained based on a deep learning model. Specifically, a first risk vector sample, a second risk vector sample, and corresponding risk level labels are pre-set, and a pre-set deep learning model is trained to obtain a risk assessment model that can output the underwriting system based on the input first and second risk vectors.

[0059] For example, a risk assessment model can determine the insurance scheme for the group to be assessed based on the input first risk vector and second risk vector. The insurance scheme includes at least: sum insured, deductible, payout ratio, and insurance coverage items.

[0060] In some implementations, the step of inputting the first risk vector and the second risk vector into a preset risk assessment model to obtain the work injury risk corresponding to the group to be assessed includes: The first risk vector and the second risk vector are input into the input layer of the risk assessment model to obtain the encoded input information; The input information is input into at least one hidden layer, and the hidden information is obtained by weighted summation and activation function corresponding to the hidden layer. The hidden information is input into the output layer to obtain the work injury risk level determined by the output layer based on the hidden information; Based on the pre-defined correspondence between work injury risk levels and insurance plans, the corresponding insurance plan for the work injury risk level is determined.

[0061] This application provides a method for assessing employee work injury risks based on an AI-powered large-scale model. The method involves: acquiring group information of the group to be assessed; constructing a first risk vector corresponding to the group based on the group information; acquiring member information of the group to be assessed; determining an assessment scheme for the group based on the member information using a preset rule engine; obtaining the assessment result of the group based on the assessment scheme; constructing a second risk vector corresponding to the group based on the assessment result; inputting the first and second risk vectors into a preset risk assessment model to obtain the work injury risk corresponding to the group to be assessed; and matching an insurance scheme for the group to be assessed based on the work injury risk. By assessing the risks of the group as a whole and the characteristics of each individual member within the group separately, and adjusting the assessment scheme according to the characteristics of the member information, a personalized insurance scheme for different groups to be assessed is determined, thus realizing personalized group insurance schemes. Furthermore, precise insurance reduces work injury risks and improves the intelligence of work injury risk management.

[0062] Please see Figure 2 , Figure 2 This is a schematic block diagram illustrating the structure of a computer device provided in an embodiment of this application. The computer device may be a server or a terminal.

[0063] like Figure 2 As shown, the computer device includes a processor, a memory, and a network interface connected via a system bus, wherein the memory may include a storage medium and internal memory.

[0064] The storage medium may store an operating system and a computer program. This computer program includes program instructions that, when executed, cause the processor to perform any AI-based large-scale model-based employee work injury risk assessment method.

[0065] The processor provides computing and control capabilities, supporting the operation of the entire computer device.

[0066] Internal memory provides an environment for the execution of computer programs stored in the storage medium. When the computer program is executed by the processor, it enables the processor to execute any employee work injury risk assessment method based on a large AI model.

[0067] This network interface is used for network communication, such as sending assigned tasks. Those skilled in the art will understand that... Figure 2 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0068] It should be understood that the processor can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among these, a general-purpose processor can be a microprocessor or any conventional processor.

[0069] In one embodiment, the processor is configured to run a computer program stored in memory to perform the following steps: Obtain the group information of the group to be evaluated, and construct a first risk vector corresponding to the group to be evaluated based on the group information; Obtain the member information of the group to be evaluated, and determine the evaluation scheme of the group to be evaluated based on the member information using a preset rule engine; The assessment results of the group to be assessed are obtained based on the assessment scheme, and a second risk vector corresponding to the group to be assessed is constructed based on the assessment results. The first risk vector and the second risk vector are input into a preset risk assessment model to obtain the work injury risk corresponding to the group to be assessed, and an insurance plan for the group to be assessed is matched according to the work injury risk.

[0070] In some implementations, obtaining the group information of the group to be evaluated and constructing a first risk vector corresponding to the group to be evaluated based on the group information includes: Capture work injury case information and perform natural language analysis on the work injury case information, and match each work injury case information with a preset industry type; Based on the work injury case information corresponding to the preset industry types, determine the industry risk vector corresponding to each preset industry type; Determine the target industry type for matching the group information from the preset industry types; Based on the industry risk vector of the target industry type, determine the first risk vector corresponding to the group to be evaluated.

[0071] In some implementations, the member information includes: member department information; before obtaining the member information of the group to be evaluated and determining the evaluation scheme of the group to be evaluated based on the member information using a preset rule engine, the method further includes: Based on the member department information, keyword matching is performed on the member departments of the group to be evaluated to obtain the occupational types corresponding to each member department; Obtain the risk base corresponding to each occupation type. The risk base includes: random risk base and cumulative risk base.

[0072] In some implementations, the member information further includes: member age information; the step of obtaining the member information of the group to be evaluated and determining the evaluation scheme of the group to be evaluated based on the member information using a preset rule engine includes: Based on the member age information and the randomness risk base, determine the randomness risk information corresponding to each group member; Based on the member age information and the cumulative risk base, determine the cumulative risk information corresponding to each group member; The evaluation scheme for each group member is determined based on the randomness risk information and the cumulative risk information.

[0073] In some implementations, determining the assessment scheme corresponding to each group member based on the randomness risk information and the cumulative risk information includes: The random risk information and cumulative risk information of each group member are normalized respectively, and the normalized random risk information is used as the horizontal axis and the normalized cumulative risk information is used as the vertical axis to determine the position of each group member in the preset coordinate system. The evaluation scheme corresponding to the group member is determined based on the position of the group member in the preset coordinate system.

[0074] In some implementations, determining the evaluation scheme corresponding to the group member based on the group member's position in the preset coordinate system includes: The assessment scheme corresponding to the group members in the first quadrant of the preset coordinate system is determined as: first questionnaire assessment and first sampling physical examination; The assessment scheme corresponding to the group members in the second quadrant of the preset coordinate system is determined as: second questionnaire assessment and first sampling physical examination; The assessment scheme corresponding to the group members in the third quadrant of the preset coordinate system is determined as: second questionnaire assessment and second sampling physical examination; The assessment scheme corresponding to the group members in the fourth quadrant of the preset coordinate system is determined as: first questionnaire assessment and second sampling physical examination. The sampling probability of the first physical examination is greater than that of the second physical examination, and the questionnaire complexity of the first questionnaire assessment is greater than that of the second questionnaire assessment.

[0075] In some embodiments, the method further includes: Obtain the questionnaire complexity corresponding to the questionnaire assessment, and determine multiple target assessment rules from each subset of assessment rules in the preset assessment rule set based on the questionnaire complexity. The subset of assessment rules includes: lifestyle habits subset, work habits subset, and mental health subset. Multiple questionnaires are generated based on the target evaluation rules. The questionnaires include: question text and answer options. The questionnaire information generated according to different target evaluation rules is randomly mixed to obtain the survey questionnaire corresponding to the questionnaire evaluation.

[0076] In some implementations, determining the random risk information corresponding to each group member based on the member's age information and the random risk base includes: in, The randomness risk information refers to the maximum age of a member and the minimum age of a member. As the first risk reduction factor, Let x be the randomness risk base, a be the member age information, b be the first preset age, and b be the second preset age.

[0077] In some implementations, determining the cumulative risk information corresponding to each group member based on the member's age information and the cumulative risk base includes: in, The cumulative risk information refers to the maximum age of the member and the minimum age of the member. This is the second risk reduction coefficient. Let x be the cumulative risk base, x be the member's age information, and α be the risk modulation index, where 0 < α < 1.

[0078] In some implementations, the step of inputting the first risk vector and the second risk vector into a preset risk assessment model to obtain the work injury risk corresponding to the group to be assessed includes: The first risk vector and the second risk vector are input into the input layer of the risk assessment model to obtain the encoded input information; The input information is input into at least one hidden layer, and the hidden information is obtained by weighted summation and activation function corresponding to the hidden layer. The hidden information is input into the output layer to obtain the work injury risk level determined by the output layer based on the hidden information; Based on the pre-defined correspondence between work injury risk levels and insurance plans, the corresponding insurance plan for the work injury risk level is determined.

[0079] It should be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of the application. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0080] It should also be understood that the term "and / or" as used in this specification and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes such combinations. It should be noted that, herein, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0081] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments. The above descriptions are merely specific implementations of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for assessing employee work-related injury risk based on an AI large-scale model, characterized in that, The method includes: Obtain the group information of the group to be evaluated, and construct a first risk vector corresponding to the group to be evaluated based on the group information; Obtain the member information of the group to be evaluated, and determine the evaluation scheme of the group to be evaluated based on the member information using a preset rule engine; The assessment results of the group to be assessed are obtained based on the assessment scheme, and a second risk vector corresponding to the group to be assessed is constructed based on the assessment results. The first risk vector and the second risk vector are input into a preset risk assessment model to obtain the work injury risk corresponding to the group to be assessed, and an insurance plan for the group to be assessed is matched according to the work injury risk.

2. The employee work injury risk assessment method based on an AI large model according to claim 1, characterized in that, The step of obtaining the group information of the group to be evaluated and constructing a first risk vector corresponding to the group to be evaluated based on the group information includes: Capture work injury case information and perform natural language analysis on the work injury case information, and match each work injury case information with a preset industry type; Based on the work injury case information corresponding to the preset industry types, determine the industry risk vector corresponding to each preset industry type; Determine the target industry type for matching the group information from the preset industry types; Based on the industry risk vector of the target industry type, determine the first risk vector corresponding to the group to be evaluated.

3. The employee work injury risk assessment method based on an AI large model according to claim 1, characterized in that, The member information includes: member department information; before obtaining the member information of the group to be evaluated and determining the evaluation scheme of the group to be evaluated based on the member information using a preset rule engine, the method further includes: Based on the member department information, keyword matching is performed on the member departments of the group to be evaluated to obtain the occupational types corresponding to each member department; Obtain the risk base corresponding to each occupation type. The risk base includes: random risk base and cumulative risk base.

4. The employee work injury risk assessment method based on an AI large model according to claim 3, characterized in that, The member information also includes: member age information; the process of obtaining the member information of the group to be evaluated and determining the evaluation scheme of the group to be evaluated based on the member information using a preset rule engine includes: Based on the member age information and the randomness risk base, determine the randomness risk information corresponding to each group member; Based on the member age information and the cumulative risk base, determine the cumulative risk information corresponding to each group member; The evaluation scheme for each group member is determined based on the randomness risk information and the cumulative risk information.

5. The employee work injury risk assessment method based on an AI large model according to claim 4, characterized in that, The step of determining the assessment scheme corresponding to each of the group members based on the randomness risk information and the cumulative risk information includes: The random risk information and cumulative risk information of each group member are normalized respectively, and the normalized random risk information is used as the horizontal axis and the normalized cumulative risk information is used as the vertical axis to determine the position of each group member in the preset coordinate system. The evaluation scheme corresponding to the group member is determined based on the position of the group member in the preset coordinate system.

6. The employee work injury risk assessment method based on an AI large model according to claim 5, characterized in that, The step of determining the evaluation scheme corresponding to the group member based on the position of the group member in the preset coordinate system includes: The assessment scheme corresponding to the group members in the first quadrant of the preset coordinate system is determined as: first questionnaire assessment and first sampling physical examination; The assessment scheme corresponding to the group members in the second quadrant of the preset coordinate system is determined as: second questionnaire assessment and first sampling physical examination; The assessment scheme corresponding to the group members in the third quadrant of the preset coordinate system is determined as: second questionnaire assessment and second sampling physical examination; The assessment scheme corresponding to the group members in the fourth quadrant of the preset coordinate system is determined as: first questionnaire assessment and second sampling physical examination. The sampling probability of the first physical examination is greater than that of the second physical examination, and the questionnaire complexity of the first questionnaire assessment is greater than that of the second questionnaire assessment.

7. The employee work injury risk assessment method based on an AI large model according to claim 1, characterized in that, The method further includes: Obtain the questionnaire complexity corresponding to the questionnaire assessment, and determine multiple target assessment rules from each subset of assessment rules in the preset assessment rule set based on the questionnaire complexity. The subset of assessment rules includes: lifestyle habits subset, work habits subset, and mental health subset. Multiple questionnaires are generated based on the target evaluation rules. The questionnaires include: question text and answer options. The questionnaire information generated according to different target evaluation rules is randomly mixed to obtain the survey questionnaire corresponding to the questionnaire evaluation.

8. The employee work injury risk assessment method based on an AI large model according to claim 4, characterized in that, The step of determining the random risk information corresponding to each group member based on the member's age information and the random risk base includes: in, The randomness risk information refers to the maximum age of a member and the minimum age of a member. As the first risk reduction factor, Let x be the randomness risk base, a be the member age information, b be the first preset age, and b be the second preset age.

9. The employee work injury risk assessment method based on an AI large model according to claim 4, characterized in that, The step of determining the cumulative risk information corresponding to each group member based on the member's age information and the cumulative risk base includes: in, The cumulative risk information refers to the maximum age of the member and the minimum age of the member. This is the second risk reduction coefficient. Let x be the cumulative risk base, x be the member's age information, and α be the risk modulation index, where 0 < α < 1.

10. The employee work injury risk assessment method based on an AI large model according to claim 1, characterized in that, The step of inputting the first risk vector and the second risk vector into a preset risk assessment model to obtain the work injury risk corresponding to the group to be assessed includes: The first risk vector and the second risk vector are input into the input layer of the risk assessment model to obtain the encoded input information; The input information is input into at least one hidden layer, and the hidden information is obtained by weighted summation and activation function corresponding to the hidden layer. The hidden information is input into the output layer to obtain the work injury risk level determined by the output layer based on the hidden information; Based on the pre-defined correspondence between work injury risk levels and insurance plans, the corresponding insurance plan for the work injury risk level is determined.

Images