Logistics personnel risk prediction and insurance management method, device and equipment and storage medium

Abstract

The present application relates to the technical field of logistics risk management, and particularly relates to a logistics personnel risk prediction and insurance management method, device, equipment and storage medium. The logistics personnel risk prediction and insurance management method comprises the following steps: obtaining historical accident data, constructing a logistics industry risk prediction model according to the historical accident data, dynamically updating the historical accident data, and updating the logistics industry risk prediction model; obtaining an insurance request of a logistics personnel, calculating the risk level of the insurance person according to the insurance request by using the logistics industry risk prediction model; generating an insurance scheme according to the pre-constructed insurance rules and the risk level of the insurance person, and outputting the insurance scheme to the insurance request initiator; dynamically obtaining the latest state information and work information of the insurance person, and updating the insurance scheme according to the latest state information and work information. The present application can generate a differentiated insurance scheme according to the risk level of each insurance person in the logistics industry.

Description

Technical Field

[0001] This invention relates to the field of logistics risk management technology, and in particular to a method, apparatus, equipment and storage medium for risk prediction and insurance management of logistics personnel. Background Technology

[0002] Currently, the logistics industry still faces numerous technical challenges in personnel management and risk control. Traditional insurance processes rely heavily on manual operations, from employee information entry to policy verification. This not only leads to low efficiency but also increases the risk of data errors due to human intervention, impacting business accuracy. In insurance management, the lack of real-time monitoring and response technologies makes it impossible to capture changes in employee status such as resignations and job adjustments in real time. This results in a disconnect between insurance coverage and actual employee circumstances, leading to delayed or missed coverage. Furthermore, the continued reliance on paper-based processes incurs high material and management costs. The lack of intelligent premium control technology also hinders timely policy cancellations in response to changes in employee status, resulting in substantial redundant premium expenditures and driving up overall operating costs.

[0003] It is evident that existing technologies still need improvement and enhancement. Summary of the Invention

[0004] In view of the shortcomings of the prior art, the purpose of this invention is to provide a method, device, equipment and storage medium for risk prediction and insurance management of logistics personnel, aiming to solve the technical problems of lack of accuracy and low efficiency in the insurance management of logistics personnel in the prior art.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: The first aspect of this invention provides a method for risk prediction and insurance management of logistics personnel, comprising the following steps: acquiring historical accident data, constructing a logistics industry risk prediction model based on the historical accident data, dynamically updating the historical accident data, and updating the logistics industry risk prediction model; acquiring insurance requests from logistics personnel, using the logistics industry risk prediction model to calculate the risk level of the insured based on the insurance request; generating an insurance plan based on pre-constructed insurance rules and the insured's risk level, and outputting the insurance plan to the initiator of the insurance request; dynamically acquiring the latest status information and work information of the insured, and updating the insurance plan based on the latest status information and work information.

[0006] Optionally, in a first implementation of the first aspect of the present invention, the step of acquiring historical accident data, constructing a logistics industry risk prediction model based on the historical accident data, dynamically updating the historical accident data, and updating the logistics industry risk prediction model specifically includes: acquiring historical accident data, which includes: accident type, accident location data, accident personnel information, and accident occurrence time; preprocessing the historical accident data to obtain a dataset; training the XGBoost algorithm with an attention mechanism using the dataset to obtain the logistics industry risk prediction model; and dynamically updating the historical accident data and updating the logistics industry risk prediction model.

[0007] Optionally, in the second implementation of the first aspect of the present invention, the preprocessing of historical accident data to obtain a dataset specifically includes: for accident type data, setting an accident severity coefficient based on the severity of the accident type; for accident location data, obtaining the accident location coordinate information based on the accident location data, dividing accident-prone road sections based on the accident location coordinate information, and setting a road section risk coefficient based on the accident frequency of accident-prone road sections; for accident personnel information, extracting the job information and work information of the accident personnel, wherein the work information includes work duration information, work time information, and work trajectory information; for accident occurrence time, statistically analyzing the accident occurrence rate for each time period, setting a time risk coefficient for each time period based on the accident occurrence rate, and setting a time risk coefficient for each accident based on the accident occurrence time; summarizing the accident type, accident severity coefficient, accident location data, accident-prone road sections, road section risk coefficient, accident personnel information, accident occurrence time, and time risk coefficient to obtain a dataset.

[0008] Optionally, in the third implementation of the first aspect of the present invention, the step of training the XGBoost algorithm with an attention mechanism using a dataset to obtain the logistics industry risk prediction model specifically includes: the XGBoost algorithm with an attention mechanism includes: a feature attention layer, an XGBoost ensemble layer, and an attention weight optimization layer; a training set is split from the dataset and input into the feature attention layer to generate weighted features; the weighted features are input into the XGBoost ensemble layer to generate preliminary prediction results; the preliminary prediction results are backpropagated through the attention weight optimization layer to update the attention weight-related parameters; iterative training is performed, the model is evaluated, and the logistics industry risk prediction model is obtained after the evaluation is passed.

[0009] Optionally, in the fourth implementation of the first aspect of the present invention, the step of obtaining the insurance application request of logistics personnel, using the logistics industry risk prediction model, and calculating the risk level of the insured based on the insurance application request, specifically includes: obtaining the insurance application request of logistics personnel; authenticating the identity of the insured based on the insurance application request to obtain the insured's information; obtaining the insured's job information, work information, and historical accident data based on the insured's information; and inputting the insured's job information, work information, and accident data into the logistics industry risk prediction model to obtain the insured's risk level.

[0010] Optionally, in the fifth implementation of the first aspect of the present invention, the step of generating an insurance plan based on pre-built insurance rules and the risk level of the insured, and outputting the insurance plan to the initiator of the insurance request, specifically includes: obtaining the job information and risk level of all logistics personnel; setting a basic insured amount based on the job information and setting an additional insured amount based on the risk level to form an insurance insured amount plan; setting a basic coverage scope based on the job information and setting an additional coverage scope based on the risk level to form an insurance coverage scope plan; and forming insurance rules with the insurance insured amount plan and the insurance coverage scope plan; obtaining the job information and risk level of the insured, generating an insurance plan based on the insurance rules, and outputting the insurance plan to the initiator of the insurance request.

[0011] Optionally, in the sixth implementation of the first aspect of the present invention, the step of dynamically acquiring the latest status information and work information of the insured, and updating the insurance plan based on the latest status information and work information, specifically includes: periodically acquiring the latest status information, latest work information, and cumulative accident data of the insured according to a preset time, wherein the latest status information includes job information and whether the insured is employed, and the latest work information includes work duration information, work time information, and work trajectory information; for resigned employees, terminating the insurance application process; for employed employees, preprocessing the latest work information and cumulative accident data to obtain preprocessed data, inputting the preprocessed data into the logistics industry risk prediction model to obtain the latest risk level of the employed employee; and updating the insurance plan according to the job information and latest risk level of the employed employee based on the insurance rules.

[0012] A second aspect of the present invention provides a risk prediction and insurance management device for logistics personnel, comprising: a model building module, used to acquire historical accident data, construct a logistics industry risk prediction model based on the historical accident data, dynamically update the historical accident data, and update the logistics industry risk prediction model; a prediction module, used to acquire insurance requests from logistics personnel, use the logistics industry risk prediction model, and calculate the risk level of the insured based on the insurance request; a scheme generation module, used to generate an insurance scheme based on pre-built insurance rules and the risk level of the insured, and output the insurance scheme to the initiator of the insurance request; and an update module, used to dynamically acquire the latest status information and work information of the insured, and update the insurance scheme based on the latest status information and work information.

[0013] Optionally, in a first implementation of the second aspect of the present invention, the model building module includes: an acquisition submodule for acquiring historical accident data, the historical accident data including: accident type, accident location data, accident personnel information, and accident occurrence time; a preprocessing submodule for preprocessing the historical accident data to obtain a dataset; a training submodule for training the XGBoost algorithm with an attention mechanism introduced using the dataset to obtain the logistics industry risk prediction model; and an update submodule for dynamically updating the historical accident data and updating the logistics industry risk prediction model.

[0014] Optionally, in a second implementation of the second aspect of the present invention, the preprocessing submodule includes: a first processing unit, used to set an accident severity coefficient based on the severity of the accident type for accident type data; a second processing unit, used to obtain the coordinate information of the accident location based on the accident location data, and divide accident-prone road sections according to the accident location coordinate information, and set a road section risk coefficient based on the accident frequency of the accident-prone road sections; a third processing unit, used to extract the job information and work information of the accident personnel, the work information including work duration information, work time information, and work trajectory information; a fourth processing unit, used to calculate the accident occurrence rate for each time period based on the accident occurrence rate, set a time risk coefficient for each time period based on the accident occurrence rate, and set a time risk coefficient for each accident based on the accident occurrence time; and a fifth processing unit, used to summarize the accident type, accident severity coefficient, accident location data, accident-prone road sections, road section risk coefficient, accident personnel information, accident occurrence time, and time risk coefficient to obtain a dataset.

[0015] Optionally, in a third implementation of the second aspect of the present invention, the training submodule includes: an input unit, used for the XGBoost algorithm with the introduced attention mechanism including: a feature attention layer, an XGBoost ensemble layer, and an attention weight optimization layer, splitting the training set from the dataset, inputting it into the feature attention layer, and generating weighted features; an output unit, used for inputting the weighted features into the XGBoost ensemble layer, generating preliminary prediction results; and an iteration unit, used for backpropagating the preliminary prediction results through the attention weight optimization layer to update the attention weight-related parameters, iteratively training, evaluating the model, and obtaining the logistics industry risk prediction model after passing the evaluation.

[0016] Optionally, in a fourth implementation of the second aspect of the present invention, the prediction module includes: a request acquisition unit, used to acquire insurance requests from logistics personnel, and to authenticate the identity of the policyholder based on the insurance requests to acquire policyholder information; an information acquisition unit, used to acquire the policyholder's job information, work information, and historical accident data based on the policyholder information; and a prediction unit, used to input the policyholder's job information, work information, and accident data into the logistics industry risk prediction model to acquire the policyholder's risk level.

[0017] Optionally, in the fifth implementation of the second aspect of the present invention, the scheme generation module includes: a coverage amount formation unit, used to obtain the job information and risk level of all logistics personnel, set a basic coverage amount according to the job information, and set an additional coverage amount according to the risk level to form an insurance coverage amount scheme; a scope formation unit, used to set a basic coverage scope according to the job information, and set an additional coverage scope according to the risk level to form an insurance coverage scope scheme, and form insurance rules with the insurance coverage amount scheme and the insurance coverage scope scheme; and a scheme formation unit, used to obtain the job information and risk level of the insured, generate an insurance scheme according to the insurance rules, and output the insurance scheme to the initiator of the insurance request.

[0018] Optionally, in a sixth implementation of the second aspect of the present invention, the updating module includes: a periodic acquisition unit, used to periodically acquire the latest status information, latest work information, and cumulative accident data of the insured according to a preset time, wherein the latest status information includes job information and whether the insured is employed, and the latest work information includes work duration information, work time information, and work trajectory information; a termination unit, used to terminate the insurance application process for departing employees; a risk updating unit, used to preprocess the latest work information and cumulative accident data for employed employees to obtain preprocessed data, and input the preprocessed data into the logistics industry risk prediction model to obtain the latest risk level of the employed employees; and a scheme updating unit, used to update the insurance scheme according to the job information and latest risk level of the employed employees and in accordance with the insurance rules. The third aspect of the present invention provides a logistics personnel risk prediction and insurance management device, including a memory and at least one processor, wherein the memory stores computer-readable instructions; the at least one processor calls the computer-readable instructions in the memory to execute the various steps of the logistics personnel risk prediction and insurance management method described above.

[0019] A fourth aspect of the present invention provides a computer-readable storage medium storing computer-readable instructions, which, when executed by a processor, implement the various steps of the logistics personnel risk prediction and insurance management method described above.

[0020] Beneficial Effects: This invention provides a method for risk prediction and insurance management of logistics personnel. The method first constructs a logistics industry risk prediction model based on historical accident data. Using this model, it calculates the risk level of each insured person based on their insurance request, automatically assessing the risk situation of each employee and selecting an insurance plan based on the risk level. Then, based on pre-constructed insurance rules and the insured person's risk level, it generates an insurance plan tailored to each insured person's risk and job position, achieving precise and differentiated insurance coverage and reducing the disconnect between insurance coverage and actual work conditions. Finally, by dynamically acquiring the latest status and work information of the insured persons, the method updates the insurance plan accordingly. This reduces unnecessary cost waste caused by personnel changes and allows for timely modification of insurance plans based on work adjustments, thus mitigating risk. Attached Figure Description

[0021] Figure 1 This is a first flowchart of a method for predicting and managing the risks of logistics personnel provided in an embodiment of the present invention; Figure 2 This is a second flowchart of the logistics personnel risk prediction and insurance management method provided in an embodiment of the present invention; Figure 3 A third flowchart of the logistics personnel risk prediction and insurance management method provided in this embodiment of the invention; Figure 4 This is a fourth flowchart of the logistics personnel risk prediction and insurance management method provided in this embodiment of the invention; Figure 5 A fifth flowchart of the logistics personnel risk prediction and insurance management method provided in this embodiment of the invention; Figure 6 The sixth flowchart of the logistics personnel risk prediction and insurance management method provided in the embodiments of the present invention; Figure 7 The seventh flowchart of the logistics personnel risk prediction and insurance management method provided in the embodiments of the present invention; Figure 8 A schematic diagram of a logistics personnel risk prediction and insurance management device provided in an embodiment of the present invention; Figure 9 Another structural schematic diagram of the logistics personnel risk prediction and insurance management device provided in an embodiment of the present invention; Figure 10 This is a schematic diagram of the structure of the logistics personnel risk prediction and insurance management equipment provided in an embodiment of the present invention. Detailed Implementation

[0022] This invention provides a method, device, equipment, and storage medium for risk prediction and insurance management of logistics personnel. Based on historical accident data from the logistics industry, this invention constructs a risk prediction model for the logistics industry. Upon receiving an insurance request, the model predicts the insured's risk level, achieving automated and standardized assessment of individual risks for logistics personnel. This assessment serves as the core basis for matching insurance plans. Then, combining pre-established standardized insurance rules, this invention customizes an exclusive insurance plan for each insured based on key dimensions such as their risk level, matching their actual work scenario and risk level. This achieves precise and differentiated insurance services, effectively solving the problem of the disconnect between insurance coverage and actual work conditions in traditional insurance. Finally, by dynamically acquiring the insured's latest status and work information, this invention updates the insurance plan in real time based on information changes. This allows for timely optimization of insurance configuration in response to personnel departures and job adjustments, avoiding cost waste caused by redundant premium expenditures. It also allows for synchronous adjustment of protection plans based on changes in work content, accurately matching risk changes and comprehensively improving the timeliness and risk control capabilities of logistics personnel insurance management.

[0023] The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0024] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 The first embodiment of the logistics personnel risk prediction and insurance management method in this invention includes: S101. Obtain historical accident data, construct a logistics industry risk prediction model based on the historical accident data, dynamically update the historical accident data, and update the logistics industry risk prediction model. S102. Obtain insurance application requests from logistics personnel, and use the aforementioned logistics industry risk prediction model to calculate the risk level of the insured based on the insurance application requests; S103. Based on the pre-built insurance rules and the risk level of the insured, generate an insurance plan and output the insurance plan to the party initiating the insurance request; S104. Dynamically obtain the latest status information and work information of the policyholder, and update the insurance plan based on the latest status information and work information.

[0025] In this embodiment, historical accident data from the logistics industry is first collected. This historical accident data includes multi-dimensional data, allowing the model to comprehensively predict the risk level of policyholders based on multiple factors. A logistics industry risk prediction model built based on real industry accident data can more realistically reflect the risk situation of industry personnel, resulting in more reliable prediction results. Furthermore, a model update trigger mechanism can be set to dynamically update historical accident data (e.g., when the amount of new accident data reaches a preset threshold, quarterly updates, or immediate updates when new accident types appear). The supplemented historical accident data is then re-inputted into the model for secondary training, optimizing the model's feature weights and prediction logic to ensure prediction accuracy. For example, when a nationwide logistics company builds its model, it first collects all logistics accident data from the past five years (including rear-end collisions involving long-distance drivers, cargo spills by warehouse sorting personnel, and injuries caused by the operation of cold chain maintenance equipment), extracts core features, and trains a basic logistics industry risk prediction model. Subsequently, it builds a data integration module to collect new accident data and personnel operation data from the company's safety management system every month. When the amount of new accident data in a single month exceeds a set value, the new data is automatically integrated into the historical database to iterate and optimize the model. For example, for events such as "urban delivery drivers are prone to sudden braking and rear-end collisions during short-distance, high-frequency deliveries," the risk weight of such behaviors is increased.

[0026] Then, after receiving insurance applications from logistics personnel, the system transforms the personnel information in the applications into feature data that the model can recognize. This data is then used by a logistics industry risk prediction model to output the insured's risk level. The model quantifies and calculates the insured's risk score based on the feature data, and then categorizes the risk score into different risk levels according to preset thresholds (e.g., low risk: 0-30 points, medium risk: 31-60 points, high risk: 61-100 points). This automates and batches risk level assessment, replacing the traditional manual risk assessment based on experience, significantly improving the efficiency of the pre-application review process. Furthermore, the model quantifies risk based on multi-dimensional features, rather than solely on job position, enabling differentiated assessments of different risks within the same job role, making risk assessments more closely aligned with the actual work situation of the personnel.

[0027] Next, the system generates an insurance plan based on pre-built insurance rules and the insured's risk level. Specifically, the insurance rules are standardized and tiered, developed by the company in conjunction with insurance companies, based on the job characteristics of different positions in the logistics industry and the probability of loss at different risk levels. These rules clearly define the corresponding insurance type, insured amount, premium, and coverage scope, allowing for flexible online adjustments. This step avoids the problem of the same plan for the same position in traditional insurance, ensuring that insurance coverage is highly matched to the insured's actual risk level and work scenario, preventing insufficient coverage for high-risk personnel and over-insurance for low-risk personnel. Furthermore, premiums are differentiated according to risk level, with low-risk personnel paying lower premiums, avoiding all employees paying high premiums uniformly and effectively reducing the company's overall insurance costs.

[0028] To address the issue of insurance coverage gaps caused by high staff turnover and frequent job changes in the logistics industry, this embodiment dynamically acquires the latest status and work information of policyholders and updates the insurance plan accordingly. When the collected information changes, such as a change in job information, or when the update ultimately affects the predicted risk level of the logistics industry risk prediction model, an update to the insurance plan is triggered.

[0029] Please see Figure 2 In the second embodiment of the logistics personnel risk prediction and insurance management method of this invention, the step of acquiring historical accident data, constructing a logistics industry risk prediction model based on the historical accident data, dynamically updating the historical accident data, and updating the logistics industry risk prediction model includes: S201. Obtain historical accident data, which includes: accident type, accident location data, accident personnel information, and accident occurrence time; S202. Preprocess historical accident data to obtain a dataset; S203. The XGBoost algorithm with an attention mechanism is trained using a dataset to obtain the logistics industry risk prediction model; S204. Dynamically update historical accident data and update the logistics industry risk prediction model.

[0030] In this embodiment, it is crucial to acquire four types of information: accident type, accident location data, accident personnel information, and accident time. Accident type is fundamental for determining which risky behaviors are likely to trigger which types of accidents and for assessing the risk and loss levels of different accidents. Accident location data reflects the relationship between risk and operational scenarios; for example, the accident risk on highways is significantly higher than indoor operations in warehouses, and the delivery risk in urban areas is higher than on suburban roads. This type of data allows the model to learn the correlation between different operational scenarios and accident probability. Accident personnel information reflects the correlation between job differences, individual operational characteristics, and risk within the logistics industry. Accident time reveals the temporal patterns of risk; for example, the accident rate surges due to fatigue driving by long-distance drivers at night, and the road accident rate is higher on rainy or snowy days in winter than in other seasons. This type of data allows the model to learn the correlation between the time dimension and accident probability, providing characteristic support for subsequent risk assessment based on the insured's operational time.

[0031] In the preprocessing of historical accident data, methods such as data cleaning, format standardization, feature extraction and annotation, and data structuring can be used to transform scattered and non-standardized raw accident data into a unified and standardized feature dataset, thereby eliminating data noise.

[0032] In terms of model construction, this embodiment adopts the XGBoost algorithm with an attention mechanism. Compared with traditional algorithms such as logistic regression and pure XGBoost, the algorithm can automatically assign higher attention weights to the key features extracted from the four types of core data (such as job type, average daily driving mileage, work scenario, and whether it is night work), allowing the model to focus more on the core features for predicting logistics risks.

[0033] Please see Figure 3 In the third embodiment of the logistics personnel risk prediction and insurance management method of this invention, the preprocessing of historical accident data to obtain a dataset specifically includes: S301. For accident type data, set an accident severity coefficient according to the severity of the accident type; S302. For accident location data, obtain the accident location coordinate information based on the accident location data, divide accident-prone road sections based on the accident location coordinate information, and set the road section risk coefficient based on the accident frequency of accident-prone road sections; S303. For accident personnel information, extract the accident personnel's job information and work information, wherein the work information includes work duration information, work time information, and work trajectory information; S304. For the time of the accident, calculate the accident rate for each time period, set the time risk coefficient for each time period based on the accident rate, and set the time risk coefficient for each accident based on the time of the accident. S305. Summarize accident types, accident severity coefficients, accident location data, accident-prone road sections, road section risk coefficients, accident personnel information, accident occurrence time, and time risk coefficients to obtain a dataset.

[0034] In this embodiment, by quantifying risk features, extracting core information, and integrating correlation patterns for different data, the resulting standardized dataset not only meets the training requirements of the XGBoost algorithm with an attention mechanism but also improves the accuracy of prediction.

[0035] The original accident types are only qualitatively categorized (e.g., rear-end collisions, cargo tipping, equipment damage), failing to reflect the actual differences in risk and loss between different accidents. By assigning specific coefficients to accidents of different severity (e.g., a coefficient of 2.0 for major injuries and 0.5 for minor collisions), the qualitative accident types are transformed into quantitative risk values, allowing the model to directly learn the correlation between accident type and the degree of risk and loss. The accident severity coefficient enables the model to clearly distinguish the characteristic differences between high-loss and low-loss accidents, preventing the model from treating minor accidents and major accidents equally, improving the model's predictive sensitivity to high-severity accidents, and enabling more accurate identification of risk characteristics that are likely to lead to major accidents in subsequent risk assessments of policyholders.

[0036] The original accident location data only consisted of geographical coordinates or textual descriptions (e.g., Highway A, Warehouse B), which could not reflect the differences in accident probability across different road sections or scenarios. By using coordinate information to divide high-risk road sections (e.g., a section of highway, a section during the evening rush hour in urban areas) and setting risk coefficients based on accident frequency (e.g., a coefficient of 1.8 for high-risk sections and 0.9 for low-risk sections), the vague geographical information was transformed into precise scenario risk values. The road section risk coefficients provided the model with fine-grained scenario characteristics. In subsequent risk assessments of policyholders, the model could match corresponding coefficients based on the policyholder's actual work road section (e.g., a high-risk section of Highway A frequently used by long-distance drivers), enabling personalized assessments of different risk levels for the same job on different work road sections.

[0037] Accident personnel information is one of the core features of logistics risk prediction. By further extracting job information (long-distance drivers, warehouse sorting, etc.), the model can clarify the basic risk differences of different jobs. Extracting work information such as working hours, working time, and work trajectory can further explore the differences in individual work characteristics behind the job (such as the risk difference between long-distance drivers who drive an average of 600km and 400km per day, and between working at night and during the day). This allows the model to learn more granular correlation patterns between individual personnel characteristics and accident occurrence.

[0038] The original accident occurrence time is only a timestamp and cannot reflect the difference in the probability of accidents occurring in different time periods. By statistically analyzing the accident occurrence rate in various time periods (night, morning rush hour, winter, holidays, etc.) and setting corresponding time risk coefficients (such as a coefficient of 1.5 after 10 pm at night and a coefficient of 1.2 on hot summer days), the single time information is transformed into time features with risk weights. This allows the model to learn the inherent laws between the time dimension and the probability of accidents, making up for the shortcomings of prediction based solely on personnel and scene features.

[0039] Please see Figure 4 In the fourth embodiment of the logistics personnel risk prediction and insurance management method of this invention, the step of training the XGBoost algorithm with an attention mechanism using a dataset to obtain the logistics industry risk prediction model specifically includes: S401. The XGBoost algorithm with an attention mechanism includes: a feature attention layer, an XGBoost ensemble layer, and an attention weight optimization layer. The training set is split from the dataset and input into the feature attention layer to generate weighted features. S402. Input the weighted features into the XGBoost ensemble layer to generate preliminary prediction results; S403. The preliminary prediction results are backpropagated through the attention weight optimization layer to update the attention weight related parameters, iteratively trained, and the model is evaluated. After the evaluation is passed, the logistics industry risk prediction model is obtained.

[0040] In this embodiment, the preprocessed standardized dataset is split into a training set, a validation set, and a test set in a ratio of 7:2:1. The training set is used for learning the core rules of the model, the validation set is used to adjust the model's hyperparameters (such as tree depth, learning rate, and feature dimension of the attention layer) during training, and the test set is used to finally verify the prediction accuracy of the model.

[0041] The preprocessed dataset contains multi-dimensional features such as personnel, scenarios, time, and accident losses (e.g., job position, road segment risk coefficient, time risk coefficient, working hours, accident severity coefficient, etc.). However, the contribution of different features to logistics risk prediction varies significantly (e.g., job position, average daily mileage, and road segment risk coefficient are core features, while the provincial administrative region where the accident occurred and the accident record number are weakly correlated or irrelevant features). The feature attention layer will autonomously learn to assign differentiated attention weights to different features, assigning high weights to core features and low weights to weakly correlated features, so that subsequent model training focuses on the key features that truly affect the occurrence of logistics accidents.

[0042] XGBoost, a classic gradient boosting tree ensemble algorithm, inherently possesses advantages in handling nonlinear features, resisting overfitting, and achieving high prediction accuracy. Inputting core features weighted by an attention layer allows XGBoost training to focus entirely on the core risk characteristics of the logistics industry, avoiding interference from irrelevant features that could weaken the algorithm's advantages. This enables the model to more efficiently learn the correlation between core features and logistics accident occurrences, generating preliminary predictions that better reflect the actual risk situation in the logistics industry. The XGBoost ensemble layer, through step-by-step training and error correction of multiple CART decision trees, weights and integrates the predictions from multiple trees to obtain the final preliminary prediction. Compared to a single decision tree model, this effectively reduces the risk of overfitting from a single tree and improves the model's adaptability to different samples in the logistics accident dataset.

[0043] Finally, the attention weight optimization layer compares the initial prediction results of the XGBoost ensemble layer with the true labels of the dataset (such as the actual accident severity coefficient and actual risk level), calculates the prediction error, and feeds the error back to the feature attention layer through the backpropagation algorithm. This dynamically updates the attention weight parameters of each feature, and multiple rounds of iterative training allow the model to fully learn the patterns of logistics risks, achieving model convergence. The model evaluation stage uses reserved validation and test sets to comprehensively evaluate the trained model. Only when the evaluation metrics reach preset thresholds will the final model be determined.

[0044] Please see Figure 5 In the fifth embodiment of the logistics personnel risk prediction and insurance management method of this invention, the step of obtaining the logistics personnel's insurance request, using the logistics industry risk prediction model, and calculating the risk level of the insured based on the insurance request, specifically includes: S501. Obtain the insurance application request from the logistics personnel, and verify the identity of the policyholder based on the insurance application request to obtain the policyholder information; S502. Obtain the policyholder's job information, work information, and historical accident data based on the policyholder's information; S503. Input the policyholder's job information, work information, and accident data into the logistics industry risk prediction model to obtain the policyholder's risk level.

[0045] In this embodiment, after obtaining the insurance application request from logistics personnel, the applicant's identity is authenticated to accurately verify their authenticity. During identity verification, a fusion of multimodal biometrics and edge computing can be used. Multimodal biometrics integrates two or more non-replicable biometric features strongly bound to the human body, categorized into physiological features (inherent and unalterable) and behavioral features, such as a combination of face, fingerprint, and voiceprint. Multi-feature cross-verification can directly identify forgery, solving the problems of impersonation and identity theft. Edge computing, on the other hand, pushes computing, storage, and data processing capabilities, originally concentrated in the cloud, down to edge nodes closer to the data collection point. Logistics personnel's biometrics are sensitive personal information; uploading them entirely to the cloud poses risks of leakage and tampering during transmission and storage. In this fusion solution, the original biometric data is stored entirely at the edge, with the cloud only synchronizing authentication results. Furthermore, edge-side data supports encrypted storage, mitigating data privacy leaks at the source.

[0046] After identity verification, the system further obtains the policyholder's information. Policyholders' self-submitted insurance requests only provide basic identity information and cannot fully include core features for model prediction, such as job details, work characteristics, and historical accident data. These data are crucial input dimensions for early model training. By using the policyholder's unique identifier to connect with various enterprise systems, the system can selectively obtain job information, work details, and historical accident data. This supplements the comprehensive feature data required for risk assessment, preventing incomplete or distorted model predictions due to missing information.

[0047] Furthermore, the acquired job information, work information, and historical accident data are completely aligned with the feature dimensions and data format used in the early model training. No additional feature conversion is required; only format adjustments are needed according to the previous steps. This allows the model to directly call upon the previously learned logistics industry risk patterns for prediction, avoiding model prediction bias caused by data dimension mismatch and fully leveraging the model's accurate prediction capabilities.

[0048] Please see Figure 6 In the sixth embodiment of the logistics personnel risk prediction and insurance management method of this invention, the step of generating an insurance plan based on pre-built insurance rules and the risk level of the insured, and outputting the insurance plan to the initiator of the insurance request, specifically includes: S601. Obtain job information and risk level of all logistics personnel, set basic insured amount based on job information, and set additional insured amount based on risk level to form an insured amount plan; S602. Set the basic coverage scope based on job information, set the additional coverage scope based on risk level, form an insurance coverage scope plan, and form insurance rules based on the insurance amount plan and the coverage scope plan; S603. Obtain the policyholder's job information and risk level, generate an insurance plan according to the insurance rules, and output the insurance plan to the party initiating the insurance request.

[0049] In this embodiment, by binding the basic insured amount to job attributes and aligning it with the core risk loss characteristics of different positions in logistics, and then adding an individual risk level associated with the insured amount, differentiated quantification of insured amounts for the same position is achieved. For example, the work scenarios and operational content of different positions in the logistics industry vary significantly, and the corresponding basic accident loss risks are also quite different (e.g., long-distance drivers face higher accident risks than warehouse sorting personnel). Those assessed as high-risk personnel receive corresponding high insured amounts, while low-risk personnel maintain the basic insured amount or receive additional low insured amounts. This allows the insured amount to be dynamically adjusted according to individual risk, ensuring that the insured amount not only matches the common risks of the position but also aligns with the individualized risks of the personnel.

[0050] Similarly, the selection of coverage is also tied to job position and risk level. The risk to high-risk individuals is not only reflected in the greater extent of potential losses but also in the more complex risk scenarios. Setting additional coverage based on risk level allows for the addition of specific risk scenario-based protection for high-risk individuals, enabling the coverage to dynamically expand according to their individual risk level.

[0051] By constructing rules from both the insured amount and the scope of coverage, a comprehensive risk-and-solution matching system is formed, enabling the insurance rules to be accurately matched with the risk characteristics of logistics personnel in all dimensions. This ensures both coverage of the loss amount and coverage of the risk scenarios.

[0052] Please see Figure 7 In the seventh embodiment of the logistics personnel risk prediction and insurance management method of this invention, the step of dynamically acquiring the latest status information and work information of the insured, and updating the insurance plan according to the latest status information and work information, specifically includes: S701. According to a preset time, periodically obtain the policyholder's latest status information, latest work information and cumulative accident data. The latest status information includes job information and whether the policyholder is employed. The latest work information includes work duration information, work time information and work trajectory information. S702. For departing employees, the insurance application process shall be terminated; S703. For employees, the latest work information and cumulative accident data are preprocessed to obtain preprocessed data, and the preprocessed data is input into the logistics industry risk prediction model to obtain the latest risk level of employees. S704. Update the insurance plan according to the job information and latest risk level of the employees, in accordance with the aforementioned insurance rules.

[0053] In this embodiment, multi-dimensional dynamic information of the insured is collected at preset intervals to ensure the timeliness of the information and provide accurate feature data for subsequent risk reassessment and plan updates. Core risk characteristics of logistics personnel, such as their job positions, work trajectories, and working hours, are constantly changing. Regularly collecting information at preset intervals (e.g., monthly, quarterly) can promptly capture these changes in risk characteristics, preventing long-term disconnect between the insurance plan and actual risks due to information lag. This ensures that risk reassessment and plan updates are always based on the insured's latest actual situation. Furthermore, including cumulative accident data after insurance coverage in the collection scope allows for complete tracking of the insured's risk behavior characteristics throughout the entire lifecycle, avoiding the one-sidedness of relying solely on pre-insurance data for risk assessment. The model can relearn the insured's risk patterns by combining post-insurance accident data, making the calculation of the latest risk level more closely match the actual risk level of their operations and improving the accuracy of risk reassessment.

[0054] To address the issue of high employee turnover in the logistics industry, this embodiment also automates and promptly terminates the insurance application process for departing employees, eliminating redundant premium payments from the source. It eliminates the need for manual follow-up by company HR, saving manpower and time costs, while also avoiding issues such as missed, incorrect, or delayed reporting in manual operations.

[0055] In response to changes in the risk characteristics of current employees, this embodiment re-invokes the risk prediction model for quantitative assessment and outputs the latest risk level. This step, by standardizing and preprocessing this latest data before inputting it into the model to recalculate the risk level, ensures that the risk assessment results fully reflect the insured's current actual operational risks, avoiding the bias of a single assessment.

[0056] The solution update in this embodiment combines job information (basic risk) and the latest risk level (individual dynamic risk), which not only ensures that the solution matches the inherent basic risk of the job, but also takes into account the dynamic risk of individuals due to changes in work characteristics and accident experience. Compared with single-dimensional adjustment, it is more accurate and more in line with the actual risk characteristics of logistics personnel.

[0057] The above describes the method for risk prediction and insurance management of logistics personnel in embodiments of the present invention. The following describes the device for risk prediction and insurance management of logistics personnel in embodiments of the present invention. Please refer to [link / reference]. Figure 8 One embodiment of the logistics personnel risk prediction and insurance management device of the present invention includes: The model building module 10 is used to acquire historical accident data, build a logistics industry risk prediction model based on the historical accident data, dynamically update the historical accident data, and update the logistics industry risk prediction model. Prediction module 20 is used to obtain insurance requests from logistics personnel and, using the logistics industry risk prediction model, calculate the risk level of the insured based on the insurance request. The scheme generation module 30 is used to generate an insurance scheme based on the pre-built insurance rules and the risk level of the insured, and output the insurance scheme to the initiator of the insurance request; The update module 40 is used to dynamically obtain the latest status information and work information of the policyholder, and update the insurance plan based on the latest status information and work information.

[0058] Please see Figure 9 One embodiment of the logistics personnel risk prediction and insurance management device of the present invention includes: The model building module 10 is used to acquire historical accident data, build a logistics industry risk prediction model based on the historical accident data, dynamically update the historical accident data, and update the logistics industry risk prediction model. Prediction module 20 is used to obtain insurance requests from logistics personnel and, using the logistics industry risk prediction model, calculate the risk level of the insured based on the insurance request. The scheme generation module 30 is used to generate an insurance scheme based on the pre-built insurance rules and the risk level of the insured, and output the insurance scheme to the initiator of the insurance request; Update module 40 is used to dynamically obtain the latest status information and work information of the policyholder, and update the insurance plan according to the latest status information and work information; In this embodiment, the model building module 10 includes: The acquisition submodule 11 is used to acquire historical accident data, which includes: accident type, accident location data, accident personnel information, and accident occurrence time; Preprocessing submodule 12 is used to preprocess historical accident data to obtain a dataset; Training submodule 13 is used to train the XGBoost algorithm with an attention mechanism using a dataset to obtain the logistics industry risk prediction model. The update submodule 14 is used to dynamically update historical accident data and update the logistics industry risk prediction model. In this embodiment, the preprocessing submodule 12 includes: The first processing unit 121 is used to set an accident severity coefficient for accident type data according to the severity of the accident type; The second processing unit 122 is used to obtain the coordinate information of the accident location based on the accident location data, divide the accident-prone road sections according to the accident location coordinate information, and set the road section risk coefficient according to the accident frequency of the accident-prone road sections. The third processing unit 123 is used to extract the job information and work information of the accident personnel from the accident personnel information. The work information includes work duration information, work time information, and work trajectory information. The fourth processing unit 124 is used to calculate the accident occurrence rate for each time period based on the accident occurrence time, set the time risk coefficient for each time period based on the accident occurrence rate, and set the time risk coefficient for each accident based on the accident occurrence time. The fifth processing unit 125 is used to summarize accident type, accident severity coefficient, accident location data, accident-prone road sections, road section risk coefficient, accident personnel information, accident occurrence time and time risk coefficient to obtain a dataset; In this embodiment, the training submodule 13 includes: Input unit 131, used for the XGBoost algorithm with introduced attention mechanism, includes: a feature attention layer, an XGBoost ensemble layer, and an attention weight optimization layer, splitting the training set from the dataset and inputting it into the feature attention layer to generate weighted features; Output unit 132 is used to input weighted features into the XGBoost ensemble layer to generate preliminary prediction results; The iteration unit 133 is used to backpropagate the preliminary prediction results through the attention weight optimization layer to update the attention weight related parameters, iterate the training, evaluate the model, and obtain the logistics industry risk prediction model after the evaluation is passed. In this embodiment, the prediction module 20 includes: The request acquisition unit 21 is used to acquire the insurance application request of logistics personnel, and to authenticate the identity of the policyholder based on the insurance application request of logistics personnel in order to obtain the policyholder information. Information acquisition unit 22 is used to acquire the policyholder's job information, work information and historical accident data based on the policyholder's information; Prediction unit 23 is used to input the insured's job information, work information and accident occurrence data into the logistics industry risk prediction model to obtain the insured's risk level; In this embodiment, the scheme generation module 30 includes: The insured amount formation unit 31 is used to obtain the job information and risk level of all logistics personnel, set the basic insured amount based on the job information, and set the additional insured amount based on the risk level to form an insured amount plan; The scope formation unit 32 is used to set the basic coverage scope based on job information, set the additional coverage scope based on risk level, form an insurance coverage scope plan, and form insurance rules based on the insurance amount plan and the coverage scope plan. The scheme formation unit 33 is used to obtain the job information and risk level of the insured, generate an insurance scheme according to the insurance rules, and output the insurance scheme to the initiator of the insurance request.

[0059] In this embodiment, the update module 40 includes: The periodic acquisition unit 41 is used to periodically acquire the policyholder's latest status information, latest work information, and cumulative accident data according to a preset time. The latest status information includes job information and whether the policyholder is employed. The latest work information includes work duration information, work time information, and work trajectory information. Suspension Unit 42 is used to suspend the insurance application process for departing employees; Risk update unit 43 is used to preprocess the latest work information and cumulative accident data of employees to obtain preprocessed data, and input the preprocessed data into the logistics industry risk prediction model to obtain the latest risk level of employees. The scheme update unit 44 is used to update the insurance scheme according to the job information and latest risk level of the employees and in accordance with the insurance rules.

[0060] The above describes the logistics personnel risk prediction and insurance management device in this embodiment of the invention from the perspective of modular functional entities. The following describes the logistics personnel risk prediction and insurance management equipment in this embodiment of the invention from the perspective of hardware processing.

[0061] The logistics personnel risk prediction and insurance management device of this invention first constructs a logistics industry risk prediction model based on historical accident data. After receiving insurance requests from logistics personnel, it automatically quantifies and calculates the risk level of each insured person through this model, completing an automated assessment of individual employee risk and providing a basis for generating differentiated insurance plans. Secondly, based on pre-built standardized insurance rules, combined with the insured person's risk level and job characteristics, it generates a precise and suitable exclusive insurance plan for each insured person, effectively solving the problem of the disconnect between insurance coverage and the actual work scenario and risk level of personnel in traditional insurance, and achieving differentiated and precise insurance. Finally, it continuously and dynamically collects the latest status and work information of the insured persons, and updates the insurance plan in real time accordingly. This not only avoids the waste of redundant premiums caused by personnel turnover and job adjustments, but also adapts to the risk fluctuations after changes in personnel's work status, comprehensively optimizing cost control and strengthening risk prevention and control capabilities.

[0062] Figure 10This is a schematic diagram of a logistics personnel risk prediction and insurance management device 900 provided in an embodiment of the present invention. The device 900 can vary significantly due to different configurations or performance characteristics. It may include one or more central processing units (CPUs) 910 (e.g., one or more processors) and a memory 920, and one or more storage media 930 (e.g., one or more mass storage devices) storing application programs 933 or data 932. The memory 920 and storage media 930 can be temporary or persistent storage. The program stored in the storage media 930 may include one or more modules (not shown in the diagram), each module including a series of instruction operations on the logistics personnel risk prediction and insurance management device 900. Furthermore, the processor 910 may be configured to communicate with the storage media 930 and execute the series of instruction operations in the storage media 930 on the logistics personnel risk prediction and insurance management device 900 to implement the steps of the logistics personnel risk prediction and insurance management methods provided in the above-described method embodiments.

[0063] The logistics personnel risk prediction and insurance management device 900 may also include one or more power supplies 940, one or more wired or wireless network interfaces 950, one or more input / output interfaces 960, and / or one or more operating systems 931, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will understand that... Figure 10 The illustrated structure of the logistics personnel risk prediction and insurance management equipment does not constitute a limitation on the equipment itself. It may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0064] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the steps of a logistics personnel risk prediction and insurance management method.

[0065] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the device or apparatus described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0066] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0067] It is understood that those skilled in the art can make equivalent substitutions or modifications to the technical solution and inventive concept of the present invention, and all such substitutions or modifications should fall within the protection scope of the appended claims.

Claims

1. A method for risk prediction and insurance management of logistics personnel, characterized in that, Includes the following steps: Acquire historical accident data, construct a logistics industry risk prediction model based on the historical accident data, dynamically update the historical accident data, and update the logistics industry risk prediction model accordingly. Obtain insurance application requests from logistics personnel, and use the aforementioned logistics industry risk prediction model to calculate the risk level of the insured based on the insurance application requests; Based on the pre-built insurance rules and the risk level of the insured, an insurance plan is generated and output to the party initiating the insurance request; The system dynamically acquires the policyholder's latest status and work information, and updates the insurance plan based on this information.

2. The method for risk prediction and insurance management of logistics personnel according to claim 1, characterized in that, The process of acquiring historical accident data, constructing a logistics industry risk prediction model based on the historical accident data, dynamically updating the historical accident data, and updating the logistics industry risk prediction model specifically includes: Acquire historical accident data, which includes: accident type, accident location data, accident personnel information, and accident occurrence time; Historical accident data is preprocessed to obtain a dataset; The XGBoost algorithm with an attention mechanism was trained using a dataset to obtain the logistics industry risk prediction model. Historical accident data is dynamically updated, and the logistics industry risk prediction model is updated accordingly.

3. The method for risk prediction and insurance management of logistics personnel according to claim 2, characterized in that, The process of preprocessing historical accident data to obtain a dataset specifically includes: For accident type data, an accident severity coefficient is set according to the severity of the accident type; For accident location data, obtain the accident location coordinate information based on the accident location data, divide accident-prone road sections based on the accident location coordinate information, and set the road section risk coefficient based on the accident frequency of accident-prone road sections; For information on personnel involved in the accident, extract their job information and work information, including work duration information, work time information, and work trajectory information. For the time of the accident, the accident rate of each time period is calculated, and a time risk coefficient is set for each time period based on the accident rate. A time risk coefficient is also set for each accident based on the time of the accident. The dataset is obtained by summarizing accident types, accident severity coefficients, accident location data, accident-prone road sections, road section risk coefficients, accident personnel information, accident occurrence time, and time risk coefficients.

4. The method for risk prediction and insurance management of logistics personnel according to claim 2, characterized in that, The method of training the XGBoost algorithm with an attention mechanism on a dataset to obtain the logistics industry risk prediction model specifically includes: The XGBoost algorithm with an attention mechanism includes: a feature attention layer, an XGBoost ensemble layer, and an attention weight optimization layer. The training set is split from the dataset and input into the feature attention layer to generate weighted features. The weighted features are input into the XGBoost ensemble layer to generate preliminary prediction results; The preliminary prediction results are backpropagated through the attention weight optimization layer to update the attention weight related parameters. The model is then iteratively trained and evaluated. Once the evaluation is passed, the logistics industry risk prediction model is obtained.

5. The method for risk prediction and insurance management of logistics personnel according to claim 1, characterized in that, The process of obtaining insurance requests from logistics personnel involves using the logistics industry risk prediction model to calculate the risk level of the insured based on the insurance request, specifically including: Obtain insurance application requests from logistics personnel, verify the identity of the policyholder based on the insurance application requests, and obtain the policyholder information; Based on the policyholder's information, obtain the policyholder's job information, work information, and historical accident data; The policyholder's job information, work information, and accident data are input into the logistics industry risk prediction model to obtain the policyholder's risk level.

6. The method for risk prediction and insurance management of logistics personnel according to claim 1, characterized in that, The process of generating an insurance plan based on pre-built insurance rules and the insured's risk level, and then outputting the insurance plan to the party initiating the insurance request, specifically includes: Obtain job information and risk level of all logistics personnel, set basic insured amount based on job information, and set additional insured amount based on risk level to form an insurance insured amount plan; The basic coverage scope is set according to the job information, and the additional coverage scope is set according to the risk level to form an insurance coverage plan. The insurance rules are formed by the insurance amount plan and the coverage scope plan. Obtain the policyholder's job information and risk level, generate an insurance plan according to the insurance rules, and output the insurance plan to the party initiating the insurance request.

7. The method for risk prediction and insurance management of logistics personnel according to claim 1, characterized in that, The process of dynamically acquiring the policyholder's latest status and work information, and updating the insurance plan based on this information, specifically includes: According to a preset time, the latest status information, latest work information and cumulative accident data of the insured are periodically obtained. The latest status information includes job information and whether the insured is employed. The latest work information includes work duration information, work time information and work trajectory information. For departing employees, the insurance application process will be suspended; For current employees, the latest work information and cumulative accident data are preprocessed to obtain preprocessed data. The preprocessed data is then input into the logistics industry risk prediction model to obtain the latest risk level of the current employees. Based on the job information and latest risk level of the employees, the insurance plan will be updated in accordance with the aforementioned insurance rules.

8. A device for risk prediction and insurance management of logistics personnel, characterized in that, include: The model building module is used to acquire historical accident data, build a logistics industry risk prediction model based on the historical accident data, dynamically update the historical accident data, and update the logistics industry risk prediction model. The prediction module is used to obtain insurance requests from logistics personnel and, using the aforementioned logistics industry risk prediction model, calculate the risk level of the insured based on the insurance request. The scheme generation module is used to generate an insurance scheme based on pre-built insurance rules and the risk level of the insured, and output the insurance scheme to the party initiating the insurance request; The update module is used to dynamically obtain the latest status and work information of the policyholder and update the insurance plan based on the latest status and work information.

9. A risk prediction and insurance management device for logistics personnel, characterized in that, The method includes a memory and at least one processor, wherein the memory stores computer-readable instructions; the at least one processor invokes the computer-readable instructions in the memory to perform various steps of the logistics personnel risk prediction and insurance management method as described in any one of claims 1-7.

10. A computer-readable storage medium storing computer-readable instructions thereon, characterized in that, When the computer-readable instructions are executed by a processor, they implement each step of the logistics personnel risk prediction and insurance management method as described in any one of claims 1-7.

Images