Method, apparatus, computer device, storage medium and program product for identifying personal risks

Abstract

The present application relates to a method, device, computer equipment, computer-readable storage medium, and computer program product for personal risk identification. The method includes: collecting multiple personal safety data information from a multi-source information system, where each personal safety data information is respectively matched with corresponding risk information; the risk information includes a risk level and a risk type; preprocessing the multiple personal safety data information to obtain multiple preprocessed data information; the preprocessed data information conforms to a data quality strategy; performing data analysis on each preprocessed data information based on data mining technology, and extracting personal safety features from each preprocessed data information; based on each personal safety feature, performing risk prediction through the constructed initial personal risk identification model to obtain prediction risk information respectively corresponding to each personal safety data information. Using this method can improve the reliability and flexibility of the power supply enterprise's personal safety management for its employees.

Description

Technical Field

[0001] This application relates to the technical fields of big data analysis, artificial intelligence, and risk management, and particularly relates to a method, device, computer device, storage medium, and program product for personal risk identification. Background Art

[0002] In today's complex and ever-changing industrial and social environment, personal safety risks have become increasingly prominent, posing severe challenges to the safe production of enterprises and institutions and the personal health of employees. Traditional risk management methods often rely on manual inspections, empirical judgments, or single data source analysis. These methods are not only inefficient but also difficult to comprehensively and accurately capture potential risk factors, resulting in the lag and incompleteness of risk prevention and control. With the rapid development of technologies such as big data and artificial intelligence, intelligent risk identification methods based on multi-source data information have gradually become a research hotspot and a new trend in practical applications.

[0003] Multi-source data information, including but not limited to production data, equipment status, environmental parameters, employee behavior records, mental health assessments, social media activities, etc., covers all aspects of employees' work and life, providing a rich data basis for comprehensively and deeply analyzing personal safety risks. However, the sources of these data are extensive, the formats are diverse, and the quality is uneven, presenting many challenges for direct application in risk identification. First, the data formats, units, dimensions, etc. between different data sources may be inconsistent, requiring data cleaning and preprocessing; second, the timeliness and integrity of the data are also important factors affecting the accuracy of risk identification; finally, how to extract useful information related to risk identification from massive data and build an effective model for prediction and judgment is the core issue of intelligent risk identification methods.

[0004] Currently, although some risk identification methods based on single data sources or simple data fusion have been proposed and applied, these methods often have the following deficiencies: first, the data sources are single and cannot comprehensively reflect the diversity of risk factors; second, the data processing methods are simple and it is difficult to fully exploit the hidden information and value in the data; third, the model construction lacks pertinence and cannot be customized according to specific risk scenarios and requirements. Summary of the Invention

[0005] Based on this, it is necessary to provide a method, device, computer device, storage medium, and program product for personal risk identification that can fuse multi-source data information and perform complex data processing to address the above technical problems.

[0006] In a first aspect, this application provides a method for personal risk identification, including:

[0007] Collect multiple personal safety data information from a multi-source information system, and each personal safety data information is respectively matched with corresponding risk information; the risk information includes a risk level and a risk type;

[0008] Preprocess the multiple personal safety data information to obtain multiple preprocessed data information; the preprocessed data information conforms to the data quality strategy;

[0009] Based on data mining techniques, perform data analysis on each preprocessed data information, and extract personal safety features from each preprocessed data information;

[0010] Based on each personal safety feature, perform risk prediction through the constructed initial personal risk identification model to obtain the predicted risk information corresponding to each personal safety data information.

[0011] In one embodiment, the method for personal risk identification further includes:

[0012] Adjust the model parameters of the initial personal risk identification model through each predicted risk information and each risk information to obtain a personal risk identification model for personal risk identification;

[0013] Deploy the personal risk identification model to the actual business system, and through the personal risk identification model, predict the real-time risk information of the employees of the power supply enterprise for the real-time personal safety data information of the employees of the power supply enterprise; the real-time risk information includes a real-time risk level and a real-time risk type; the real-time personal safety data information includes at least the working environment and behavior patterns;

[0014] Determine the risk warning matching the real-time risk information, and determine the intervention measures matching the real-time risk information.

[0015] In one embodiment, determining the risk warning matching the real-time risk information includes:

[0016] Based on the real-time risk level and the real-time risk type, determine the potential risks of the employees of the power supply enterprise;

[0017] Trigger a risk warning when the potential risks of the employees of the power supply enterprise reach a preset threshold.

[0018] In one embodiment, determining the intervention measures matching the real-time risk information includes:

[0019] Determine the intervention strategies corresponding to each real-time risk type;

[0020] Establish a feedback opinion collection mechanism according to the real-time risk type in the real-time risk information, and collect feedback opinions for different real-time risk types based on the feedback opinion collection mechanism;

[0021] Optimize and adjust the intervention strategies for real-time risk types based on the feedback of each real-time risk type.

[0022] In one embodiment, the initial personal risk identification model is constructed as follows:

[0023] Based on each personal safety feature, construct the corresponding unverified personal risk identification models to be verified respectively through multiple model construction algorithms; the multiple model construction algorithms include at least one of logistic regression algorithm, decision tree algorithm, random forest algorithm, gradient boosting tree algorithm, and neural network algorithm;

[0024] Based on each personal safety feature, perform cross-validation on each unverified personal risk identification model to obtain the model performance corresponding to each unverified personal risk identification model;

[0025] Determine the unverified personal risk identification model with the optimal model performance as the initial personal risk identification model.

[0026] In one embodiment, perform data analysis on each preprocessed data information based on data mining technology, and extract personal safety features from each preprocessed data information, including:

[0027] Perform data analysis on each preprocessed data information by using any one of correlation analysis, principal component analysis, and mutual information method to obtain the data analysis result;

[0028] Based on the data analysis result, screen out the feature subset with the highest correlation and the largest amount of information related to personal safety from each preprocessed data information, and determine the feature subset as the personal safety feature;

[0029] Based on each personal safety feature, perform risk prediction through the constructed initial personal risk identification model to obtain the predicted risk information corresponding to each personal safety data information, including:

[0030] Use at least one of clustering analysis and association rule mining technology through the constructed initial personal risk identification model to identify the potential patterns and association relationships in each personal safety feature;

[0031] Adopt at least one of time series analysis method and machine learning regression model method through the initial personal risk identification model, and predict the predicted risk information corresponding to each personal safety data information according to the collected potential patterns and association relationships.

[0032] In a second aspect, the present application also provides a personal risk identification device, including:

[0033] A data information collection module, which is used to collect multiple personal safety data information from a multi-source information system, and each personal safety data information is respectively matched with corresponding risk information; the risk information includes a risk level and a risk type;

[0034] A data information processing module, which is used to preprocess multiple personal safety data information to obtain multiple preprocessed data information; the preprocessed data information conforms to the data quality strategy;

[0035] A data analysis module, which is used to perform data analysis on each preprocessed data information based on data mining technology, and extract personal safety features from each preprocessed data information;

[0036] A model prediction module, which is used to perform risk prediction based on each personal safety feature through the constructed initial personal risk identification model, and obtain the predicted risk information corresponding to each personal safety data information.

[0037] In a third aspect, the present application also provides a computer device, including a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0038] Collect multiple personal safety data information from a multi-source information system, and each personal safety data information is respectively matched with corresponding risk information; the risk information includes a risk level and a risk type;

[0039] Preprocess multiple personal safety data information to obtain multiple preprocessed data information; the preprocessed data information conforms to the data quality strategy;

[0040] Perform data analysis on each preprocessed data information based on data mining technology, and extract personal safety features from each preprocessed data information;

[0041] Perform risk prediction based on each personal safety feature through the constructed initial personal risk identification model, and obtain the predicted risk information corresponding to each personal safety data information.

[0042] In a fourth aspect, the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0043] Collect multiple personal safety data information from a multi-source information system, and each personal safety data information is respectively matched with corresponding risk information; the risk information includes a risk level and a risk type;

[0044] Preprocess multiple personal safety data information to obtain multiple preprocessed data information; the preprocessed data information conforms to the data quality strategy;

[0045] Perform data analysis on each pre - processed data information based on data mining technology, and extract personal safety features from each pre - processed data information;

[0046] Based on each personal safety feature, perform risk prediction through the constructed initial personal risk identification model to obtain the predicted risk information corresponding to each personal safety data information.

[0047] In a fifth aspect, the present application also provides a computer program product, including a computer program, which when executed by a processor implements the following steps:

[0048] Collect multiple personal safety data information from a multi - source information system, and each personal safety data information is respectively matched with a corresponding risk information; the risk information includes a risk level and a risk type;

[0049] Pre - process the multiple personal safety data information to obtain multiple pre - processed data information; the pre - processed data information conforms to the data quality strategy;

[0050] Perform data analysis on each pre - processed data information based on data mining technology, and extract personal safety features from each pre - processed data information;

[0051] Based on each personal safety feature, perform risk prediction through the constructed initial personal risk identification model to obtain the predicted risk information corresponding to each personal safety data information.

[0052] The above-mentioned methods, devices, computer equipment, computer-readable storage media, and computer program products for personal risk identification collect multiple personal safety data information from multi-source information systems, and each personal safety data information is respectively matched with corresponding risk information; the risk information includes risk levels and risk types; preprocess the multiple personal safety data information to obtain multiple preprocessed data information; the preprocessed data information conforms to the data quality strategy; perform data analysis on each preprocessed data information based on data mining technology, and extract personal safety features from each preprocessed data information; based on each personal safety feature, perform risk prediction through the constructed initial personal risk identification model to obtain the predicted risk information corresponding to each personal safety data information. First, by introducing artificial intelligence and machine learning technologies, the intelligent identification, early warning, and management of personal safety risks are realized, which can not only significantly improve the efficiency and accuracy of the safety management of power supply enterprises, but also greatly reduce the subjectivity and uncertainty of human judgment, making the safety management more scientific and accurate. Second, multi-dimensional information such as employees' personal characteristics, working environments, and historical records is fully considered, and personalized risk assessments and early warnings can be generated for different employees and different scenarios, further improving the overall safety management level. It can be seen that the reliability and flexibility of the personal safety management of power supply enterprises for their employees can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] To more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for the description of the embodiments of the present application or related technologies. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0054] Figure 1 It is an application environment diagram of the method for personal risk identification in an embodiment;

[0055] Figure 2 It is a flowchart of the method for personal risk identification in an embodiment;

[0056] Figure 3 It is a flowchart of the method for personal risk identification in another embodiment;

[0057] Figure 4 It is an overall flowchart of the method for personal risk identification based on multi-source data information in an embodiment;

[0058] Figure 5 It is a flowchart of the implementation of the personal risk identification model of the method for personal risk identification based on multi-source data information in an embodiment;

[0059] Figure 6 It is a structural block diagram of a personal risk identification device in an embodiment;

[0060] Figure 7 It is an internal structure diagram of a computer device in an embodiment. Specific implementation manners

[0061] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0062] The method for personal risk identification provided by the embodiments of the present application can be applied to an application environment as Figure 1 shown. Among them, the terminal 102 communicates with the server 104 through a network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or can be placed in the cloud or other network servers.

[0063] Specifically, taking the application to the server 104 as an example for description, multiple personal safety data information is collected from a multi-source information system. Each personal safety data information is matched with corresponding risk information, and the risk information includes a risk level and a risk type. Then, the multiple personal safety data information is preprocessed to obtain multiple preprocessed data information, and the preprocessed data information conforms to the data quality policy. Thus, based on data mining technology, data analysis is performed on each preprocessed data information, personal safety features are extracted from each preprocessed data information, and based on each personal safety feature, risk prediction is performed through the constructed initial personal risk identification model to obtain the predicted risk information corresponding to each personal safety data information. By reducing the subjectivity and uncertainty of human judgment, the safety management becomes more scientific and accurate. And multi-dimensional information such as employee personal characteristics, working environment, historical records, etc. are fully considered, further improving the overall safety management level. It can be seen from this that the reliability and flexibility of the personal safety management of power supply enterprises for power supply enterprise employees can be improved.

[0064] Among them, the terminal 102 can be, but is not limited to, various personal computers, laptop computers, smart phones, tablet computers, Internet of Things devices, and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, projection devices, etc. The portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The head-mounted devices can be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. The server 104 can be an independent physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0065] In an exemplary embodiment, as Figure 2 shown, a method for identifying personal risks is provided. In this embodiment, taking this method applied to Figure 1 the server 104 shown as an example, it can be understood that this method can also be applied to Figure 1 the terminal 102 shown, and can also be applied to a system including the terminal 102 and the server 104, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0066] Step 202, collect multiple personal safety data information from a multi-source information system, and each personal safety data information is respectively matched with corresponding risk information; the risk information includes a risk level and a risk type.

[0067] Among them, the multi-source information system includes an external system and each business system belonging to the power supply enterprise. The external system is a business system that does not belong to the power supply enterprise. Therefore, the external system includes, but is not limited to, a meteorological information system, a traffic management system, and a public safety information platform. Secondly, each business system belonging to the power supply enterprise includes, but is not limited to, a production management system, a human resources management system, and a safety management system.

[0068] Secondly, the aforementioned personal safety data information includes, but is not limited to: employee basic information, historical accident records, equipment status monitoring data, work environment monitoring data, weather condition data, traffic condition data, public safety event data, as well as work load and stress assessment data, safety training and qualification certification information data, special operation permit and approval data, personal health monitoring data, emergency response and rescue ability data, social network and sentiment analysis data, external threat and risk intelligence data.

[0069] The aforementioned workload and stress assessment data collect data on employees' work stress, workload, and mental health status through questionnaire surveys and heart rate monitoring means to evaluate the potential indirect impact on personal safety. The aforementioned safety training and qualification certification information data analyze the impact on employees' safety awareness and mastery of operation specifications by investigating the records of safety training courses participated by employees, obtained qualification certificates, and examination results, so as to evaluate potential risks. The aforementioned special operation permit and approval data collect monitoring data during the operation permit application, approval process, and implementation process for high-risk operations, including high-altitude operations and live-line operations, to ensure operation compliance and reduce risks.

[0070] The aforementioned personal health monitoring data collect employees' physical health indicators, including blood pressure, blood sugar, and heart rate variability, through wearable devices and regular physical examinations to evaluate the impact of their physical status on work safety. The aforementioned emergency response and rescue capacity data evaluate the completeness of the enterprise's emergency response system, including the determination of emergency plans, drill records, training and equipment conditions of rescue teams, and the efficiency and effectiveness of historical rescue operations, providing data support for quickly responding to emergencies. The aforementioned social network and sentiment analysis data use natural language processing technology to analyze employees' remarks on social media, identify negative emotions and potential psychological stress, and intervene in a timely manner to avoid work mistakes and safety accidents caused by personal emotional problems. The aforementioned external threat and risk intelligence data collect and analyze external threat information from the Internet, media, and other channels regarding influenza spread, natural disasters, and special events that may affect enterprise operations and employee safety, and make preparations in advance.

[0071] Specifically, the server collects personal safety data information from a multi-source information system, that is, data collection is carried out from the meteorological information system, traffic management system, public security information platform, production management system, human resources management system, and safety management system to obtain personal safety data information. And the collected personal safety data information is all matched with corresponding risk information. The aforementioned risk information includes risk levels and risk types. The risk levels at least include high risk, medium risk, and low risk, and the risk types at least include electric shock risk, high-altitude fall risk, mechanical injury risk, fire and explosion risk, chemical leakage risk, radiation injury risk, infectious disease risk, human error risk, natural disaster risk, violent incident risk, and psychological problem risk, etc. The risk information matched with the personal safety data information can be obtained through artificial risk assessment of each personal safety data information after information collection.

[0072] Step 204, perform preprocessing on multiple personal safety data information to obtain multiple preprocessed data information; the preprocessed data information conforms to the data quality strategy.

[0073] Among them, the preprocessed data information conforms to the data quality strategy, and the aforementioned data quality strategy at least includes: unified format, no duplicate data, no missing values, data outliers, etc. Specifically, the server preprocesses multiple personal safety data information to ensure that each preprocessed data information obtained conforms to the data quality strategy.

[0074] Since the data quality strategy at least includes: unified format, no duplicate data, no missing values, data outliers, etc., then in the process of preprocessing personal safety data information, corresponding processing needs to be carried out based on the aforementioned data quality strategy. Based on this, in a specific embodiment, the preprocessing methods include at least one of the following: format standardization, data cleaning, data deduplication, missing value processing, outlier correction, standardization processing, data integration.

[0075] The following is a detailed introduction to the aforementioned preprocessing methods:

[0076] The aforementioned format standardization is to convert various types of data into a unified format recognizable by the system, including date format, timestamp, and unified specification of currency units, that is, to convert each personal safety data information into a unified format recognizable by the system. The aforementioned data cleaning is to remove and correct incorrect, incomplete, and redundant records, that is, to remove and correct incorrect, incomplete, and redundant records in personal safety data information to reduce the interference of noise on the analysis results.

[0077] The aforementioned data deduplication is to delete duplicate records by comparing the unique identifiers in the data, that is, to match the unique identifiers in each personal safety data information to delete the personal safety data information with duplicate identifiers and avoid causing deviations in subsequent analysis. The aforementioned missing value processing adopts interpolation method, mean substitution, and filling method based on business rules, that is, to perform the aforementioned missing value processing on personal safety data information by using interpolation method, mean substitution, and filling method based on business rules to ensure the integrity of the data.

[0078] The aforementioned outlier correction is to identify and adjust those values that deviate extremely from the normal range according to the data distribution law and business logic, that is, to perform outlier identification on personal safety data information according to the data distribution law and business logic in the power supply field, and then adjust the values in personal safety data information that deviate from the normal range through the outlier identification results to make them more in line with the actual situation.

[0079] The foregoing standardization process scales the data proportionally, that is, scales each personal safety data information proportionally so that it falls into a small specific interval for unified comparison and analysis. Finally, the foregoing data integration integrates the preprocessed data into a unified data warehouse, that is, integrates the personal safety data information after the foregoing processing into a unified data warehouse to provide comprehensive data support for subsequent data mining and model training.

[0080] Step 206: Perform data analysis on each preprocessed data information based on data mining technology, and extract personal safety features from each preprocessed data information.

[0081] Specifically, the server performs data analysis on each preprocessed data information based on data mining technology, that is, through the use of various algorithms and models, performs feature selection on the preprocessed data information to extract personal safety features from each preprocessed data information. The methods of feature selection in data analysis at least include correlation analysis, principal component analysis, and mutual information method. The following is a detailed introduction to this:

[0082] In a specific embodiment, performing data analysis on each preprocessed data information based on data mining technology and extracting personal safety features from each preprocessed data information includes: performing data analysis on each preprocessed data information by using any one of correlation analysis, principal component analysis, and mutual information method to obtain a data analysis result; based on the data analysis result, screening out the feature subset with the highest correlation with personal safety and the largest amount of information from each preprocessed data information, and determining the feature subset as the personal safety feature. Specifically, the server needs to use any one of the methods of correlation analysis, principal component analysis (PCA), and mutual information (MI) method during the process of feature selection to perform data analysis on each preprocessed data information, screen out the feature subset with the highest correlation with personal safety and the largest amount of information from each preprocessed data information based on the data analysis result, and then determine the feature subset as the personal safety feature to complete feature screening.

[0083] Step 208: Based on each personal safety feature, perform risk prediction through the constructed initial personal risk identification model to obtain the predicted risk information corresponding to each personal safety data information.

[0084] Among them, the predicted risk information includes the predicted risk level and the predicted risk type. Specifically, based on each personal safety feature, risk prediction is performed through the constructed initial personal risk identification model to obtain the predicted risk information corresponding to each personal safety data information. That is, each personal safety feature is input into the pre-constructed initial personal risk identification model, and the initial personal risk identification model then performs risk prediction and outputs the predicted risk level and the predicted risk type corresponding to each personal safety data information. The aforementioned predicted risk level and predicted risk type are the predicted risk information.

[0085] The following will detail how to construct the initial personal risk identification model: In a specific embodiment, the construction method of the initial personal risk identification model is as follows: Based on each personal safety feature, the matching personal risk identification models to be verified are respectively constructed through multiple model construction algorithms; the multiple model construction algorithms include at least one of the logistic regression algorithm, the decision tree algorithm, the random forest algorithm, the gradient boosting tree algorithm, and the neural network algorithm; cross-validation is performed on each personal risk identification model to be verified to obtain the model performance of each personal risk identification model to be verified; the personal risk identification model to be verified with the optimal model performance is determined as the initial personal risk identification model.

[0086] That is, at least data feature engineering and algorithm selection and model construction are required to construct the initial personal risk identification model. The following is an introduction to this:

[0087] 1. Data feature engineering: According to the requirements of the personal safety risk assessment of power supply enterprise employees, further feature engineering is performed on the processed data, including feature construction, feature scaling, and encoding conversion, to enhance the interpretability of the model and improve the learning efficiency of the model. Specifically, categorical variables are converted into numerical features through one-hot encoding and label encoding, continuous features are normalized, and composite features are constructed according to business logic.

[0088] 2. Algorithm Selection and Model Construction: That is, the server constructs the corresponding risk identification models for the personal risks to be verified through logistic regression, decision tree, random forest, gradient boosting decision tree, and neural network algorithms respectively, and performs cross-validation on each risk identification model for the personal risks to be verified to obtain the model performance of each risk identification model for the personal risks to be verified. At this time, the server needs to select logistic regression, decision tree, random forest, gradient boosting decision tree (GBDT), and neural network algorithms to construct the personal risk identification model based on the complexity and diversity of multi-source data information, as well as the real-time and accuracy requirements of personal risk identification; in the model construction stage, cross-validation is first performed to evaluate the model performance under different algorithms and parameter combinations, and the optimal model configuration is selected. By adjusting the hyperparameters of the model (depth of the tree, learning rate, regularization coefficient), the prediction accuracy and generalization ability of the model are further optimized.

[0089] The following will detail how to do it: In a specific embodiment, based on each personal safety feature, risk prediction is performed through the constructed initial personal risk identification model to obtain the predicted risk information corresponding to each personal safety data information, including: using at least one of clustering analysis and association rule mining techniques through the constructed initial personal risk identification model to identify potential patterns and association relationships in each personal safety feature; using at least one of time series analysis methods and machine learning regression model methods through the initial personal risk identification model to predict the predicted risk information corresponding to each personal safety data information according to the collected potential patterns and association relationships.

[0090] That is, the initial personal risk identification model constructed by the server performs model identification and trend prediction. Model identification is that the initial personal risk identification model uses at least one of clustering analysis and association rule mining techniques to identify potential patterns and association relationships in each personal safety feature. Then trend prediction is performed, that is, using at least one of time series analysis methods and machine learning regression model methods, and according to the potential patterns and association relationships obtained from model identification, the predicted risk information corresponding to each personal safety data information is predicted. The aforementioned predicted risk information can be the risk information within a preset future time period.

[0091] Secondly, the risk information within a preset future time period may include, but is not limited to, electric shock, falling from height, mechanical injury, fire and explosion, chemical leakage, radiation injury, traffic accident, occupational disease, infectious disease, human error, natural disaster, violent event, biosafety threat, health problems caused by excessive psychological stress, food poisoning, etc.

[0092] It should be understood that the corresponding examples in the embodiments of the present application are all used to understand the present solution, but should not be construed as specific limitations on the present solution.

[0093] In the above method for identifying personal risks, by introducing artificial intelligence and machine learning technologies, the intelligent identification, early warning, and management of personal safety risks are realized. This can not only significantly improve the efficiency and accuracy of the safety management of power supply enterprises but also greatly reduce the subjectivity and uncertainty of human judgment, making the safety management more scientific and precise. Secondly, multi-dimensional information such as employees' personal characteristics, working environments, and historical records is fully considered, enabling the generation of personalized risk assessments and early warnings for different employees and different scenarios, further improving the overall safety management level. It can be seen that the reliability and flexibility of the personal safety management of power supply enterprises for their employees can be improved.

[0094] In an exemplary embodiment, as Figure 3 shown, the method for identifying personal risks further includes:

[0095] Step 302: Adjust the model parameters of the initial personal risk identification model through each predicted risk information and each risk information to obtain a personal risk identification model for personal risk identification.

[0096] Specifically, the server adjusts the model parameters of the initial personal risk identification model through each predicted risk information and each risk information to obtain a personal risk identification model for personal risk identification. That is, the server needs to at least perform model training and optimization on the initial personal risk identification model. The following introduces model training and optimization: After determining the initial personal risk identification model, the server uses the preprocessed and feature-engineered dataset to train the model. During the training process, appropriate loss functions (cross-entropy loss, mean squared error) are used to measure the difference between the model prediction value and the actual value, and the model parameters are updated through the gradient descent optimization algorithm to minimize the loss function.

[0097] It can be seen that the server uses appropriate loss functions (such as cross-entropy loss, mean squared error) to calculate the loss function for each predicted risk information and each risk information to obtain a loss value. At this time, the loss value can be a cross-entropy loss value or a mean squared error loss value. The aforementioned loss value is the difference between the prediction value and the actual value. Then, based on the aforementioned loss value, the model parameters of the initial personal risk identification model are adjusted through the gradient descent optimization algorithm to minimize the loss function, thereby obtaining a personal risk identification model for personal risk identification.

[0098] It can be understood that in practical applications, after adjusting the model parameters of the initial personal risk identification model, model evaluation and interpretation, as well as model deployment and monitoring, can also be performed. The following details this:

[0099] 1. Model Evaluation and Interpretation: After adjusting the model parameters of the initial personal risk identification model, an independent test set is used to evaluate the model to ensure that the model has good generalization ability, that is, it can maintain stable prediction performance on unknown data. When evaluating the model, not only common indicators such as accuracy, recall, and F1-score are concerned, but also the interpretability, stability, and real-time performance of the model are combined. Specifically, a model with strong interpretability helps the enterprise understand the internal mechanism of risk factors and determine targeted prevention and control measures; a model with good stability can maintain consistent prediction performance at different time points and data distributions; real-time performance requires the model to complete the prediction task within a short time to meet the enterprise's need to quickly respond to risk changes.

[0100] 2. Model Deployment and Monitoring: Deploy the trained and evaluated personal risk identification model to the actual business system, integrate it with the production management system, human resource management system, and safety management system of the power supply enterprise to achieve real-time monitoring and analysis. During the operation of the model, by setting up a monitoring mechanism, regularly check the prediction accuracy and stability of the model, promptly discover and handle model degradation problems. At the same time, according to business requirements and data changes, regularly update and optimize the model to maintain its best performance.

[0101] Step 304: Deploy the personal risk identification model to the actual business system, and predict the real-time risk information of power supply enterprise employees based on the real-time personal safety data information of power supply enterprise employees through the personal risk identification model; the real-time risk information includes real-time risk level and real-time risk type; the real-time personal safety data information includes at least the working environment and behavior pattern.

[0102] Among them, the real-time risk information includes real-time risk level and real-time risk type. Similar to the embodiments described above, the real-time risk level includes at least high risk, medium risk, and low risk, and the real-time risk type includes at least electric shock risk, high-altitude fall risk, mechanical injury risk, fire and explosion risk, chemical leakage risk, radiation injury risk, infectious disease risk, human error risk, natural disaster risk, violent incident risk, and psychological problem risk, etc.

[0103] Specifically, deploy the personal risk identification model to the actual business system, so as to obtain the working environment and behavior pattern of power supply enterprise employees in the actual business system, and input the working environment and behavior pattern of power supply enterprise employees into the personal risk identification model, so as to output the real-time risk information of power supply enterprise employees through the personal risk identification model. The working environment and behavior pattern of power supply enterprise employees in the actual business system belong to the personal safety data information introduced in the foregoing embodiments, that is, the working environment can be working environment monitoring data, and the behavior pattern can be work load and pressure assessment data, as well as personal health monitoring data.

[0104] Step 306: Determine the risk warning that matches the real-time risk information and determine the intervention measures that match the real-time risk information.

[0105] Specifically, determine the risk warning that matches the real-time risk information and determine the intervention measures that match the real-time risk information. As can be seen from the foregoing embodiments, the real-time risk information includes the real-time risk level and the real-time risk type. Therefore, at this time, the matching risk warning can be determined based on the real-time risk level and the real-time risk type, and the matching intervention measures can be determined based on the real-time risk type in the real-time risk information, so as to reduce the risk of personal safety accidents of employees in the power supply enterprise.

[0106] The following separately introduces the methods for determining the risk warning and determining the intervention policy. First, introduce the method for determining the risk warning that matches the real-time risk information: In an alternative embodiment, determining the risk warning that matches the real-time risk information includes: based on the real-time risk level and the real-time risk type, determining the potential risks of employees in the power supply enterprise; triggering a risk warning when the potential risks of employees in the power supply enterprise reach a preset threshold.

[0107] Among them, the risk warning is pushed in at least one of the ways of text message, email, and mobile application. That is, the risk warning is pushed through text message, email, and mobile application to ensure that relevant management personnel and employees can obtain risk information in a timely manner. Specifically, based on the real-time risk level and the real-time risk type, when the potential risks of employees in the power supply enterprise reach a preset threshold, a risk warning is triggered to ensure that management personnel and employees in the power supply enterprise can obtain risk information in a timely manner.

[0108] In an alternative embodiment, determining the intervention measures that match the real-time risk information includes: determining the intervention strategies corresponding to each real-time risk type; establishing a feedback opinion collection mechanism according to the real-time risk type in the real-time risk information, and collecting the feedback opinions for different real-time risk types respectively based on the feedback opinion collection mechanism; optimizing and adjusting the intervention strategies for the real-time risk type through the feedback opinions for different real-time risk types respectively.

[0109] As can be seen from the foregoing introduction, the real-time risk types at least include electric shock risk, high-altitude fall risk, mechanical injury risk, fire and explosion risk, chemical leakage risk, radiation injury risk, infectious disease risk, human error risk, natural disaster risk, violent incident risk, and psychological problem risk, etc. Then, corresponding intervention measures are determined and implemented for the real-time risk types, and a feedback opinion collection mechanism is established to collect the feedback opinions of employees on the intervention measures, that is, the feedback opinions for different real-time risk types are collected respectively based on the feedback opinion collection mechanism, and the intervention strategies for the real-time risk types are optimized and adjusted, so as to continuously optimize and adjust the intervention strategies and ensure the effectiveness and adaptability of the intervention measures.

[0110] The corresponding intervention measures are introduced for different real-time risk types as follows:

[0111] In the case where the real-time risk type is electric shock risk, the matched intervention measures can be to strengthen on-site safety inspections to ensure the integrity of electrical equipment and the safe operation of employees.

[0112] In the case where the real-time risk type is high-altitude fall risk, the matched intervention measures can be to strengthen the use of safety belts and the setting of safety nets.

[0113] In the case where the real-time risk type is mechanical injury risk, the matched intervention measures can be to regularly maintain and inspect mechanical equipment to ensure its safety performance, and provide safety training to operators to improve their safety awareness and operation skills.

[0114] In the case where the real-time risk type is fire and explosion risk, the matched intervention measures can be to strengthen the construction and maintenance of fire-fighting facilities to ensure unobstructed fire-fighting channels, and conduct regular fire drills to improve employees' fire emergency handling capabilities.

[0115] In the case where the real-time risk type is chemical leakage risk, the matched intervention measures can be to strictly manage the storage and use areas of chemical substances to ensure compliance with safety regulations, and equip employees with necessary protective equipment such as chemical protective clothing and gas masks.

[0116] In the case where the real-time risk type is radiation injury risk, the matched intervention measures can be to strictly control radiation sources to ensure their use under safe shielding conditions, and conduct regular health checks on employees who may be exposed to radiation.

[0117] In the case where the real-time risk type is infectious disease risk, the matched intervention measures can be to strengthen the hygiene management of the workplace, conduct regular disinfection, and encourage employees to develop good personal hygiene habits such as washing hands frequently and wearing masks.

[0118] In the case where the real-time risk type is human error risk, the matched intervention measures can be to reduce it by optimizing work processes and improving the quality of employee training. At the same time, an intelligent monitoring system is introduced to monitor abnormal behaviors in the work process in real time and intervene in a timely manner.

[0119] In the case where the real-time risk type is natural disaster risk, the matched intervention measures can be to establish a complete emergency plan, provide training on natural disaster response to employees, and conduct regular drills to ensure that they can respond quickly and effectively when natural disasters occur.

[0120] In the case where the real-time risk type is a violent incident risk, the matching intervention measures can be to strengthen the safety monitoring in the workplace, improve the safety awareness of employees, and establish an effective reporting and response mechanism to ensure that violent incidents can be handled and intervened in a timely manner when they occur.

[0121] In the case where the real-time risk type is a psychological problem risk, the matching intervention measures can be to provide psychological counseling and stress management workshops.

[0122] In this embodiment, enhancing employees' safety awareness and self-protection ability: Through the implementation of risk early warning and intervention measures, the present invention can not only effectively prevent risks before they occur, but also quickly respond when risks occur, reducing losses. Secondly, through continuous feedback and adjustment, employees can gradually enhance their safety awareness, improve their self-protection ability, and further adjust strategies to ensure the effectiveness of prevention, forming a good safety culture atmosphere.

[0123] The following introduces the complete implementation method of the personal risk identification method based on multi-source data information. In an exemplary embodiment, as Figure 4 shown, a personal risk identification method based on multi-source data information is provided. In this embodiment, this method is exemplified by being applied to Figure 1 the server 104 shown. It can be understood that this method can also be applied to Figure 1 the terminal 102 shown, and can also be applied to a system including the terminal 102 and the server 104, and is realized through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0124] Step 401, collect multiple personal safety data information from a multi-source information system, thereby realizing the integration of internal and external multi-source data.

[0125] Step 402, preprocess the multiple personal safety data information to obtain multiple preprocessed data information, thereby ensuring the data quality and providing an accurate basis for subsequent analysis.

[0126] Step 403, perform data analysis on each preprocessed data information based on data mining technology, and extract personal safety features from each preprocessed data information.

[0127] Step 404, construct an initial personal risk identification model, and use machine learning algorithms to train and learn the personal safety features to form a personal risk identification model for personal risk identification.

[0128] Step 405: Deploy the personal risk identification model to the actual business system, and predict the real-time risk information of power supply enterprise employees based on the real-time personal safety data information of power supply enterprise employees, which at least includes the working environment and behavior patterns, so as to facilitate the timely discovery of potential personal safety risks.

[0129] Step 406: Determine the risk warning that matches the real-time risk information, and determine the intervention measures that match the real-time risk information, thereby reducing the risk of personal safety accidents.

[0130] The aforementioned multi-source information system includes external systems and each business system belonging to the power supply enterprise. The external system is a business system that does not belong to the power supply enterprise. Each business system belonging to the power supply enterprise includes but is not limited to the production management system, human resource management system, and safety management system, while the external system includes but is not limited to the meteorological information system, traffic management system, and public safety information platform. The personal safety data information includes but is not limited to employee basic information, historical accident records, equipment status monitoring data, working environment monitoring data, weather condition data, traffic condition data, public safety event data, as well as work load and stress assessment data, safety training and qualification certification information data, special operation permit and approval data, personal health monitoring data, emergency response and rescue ability data, social network and sentiment analysis data, and external threat and risk intelligence data.

[0131] In the embodiment of the present application, a comprehensive and dynamic personal risk identification system is constructed by integrating a wide range of data sources inside and outside the power supply enterprise. This system not only covers internal information such as employees' personal characteristics and historical safety records, but also incorporates external environmental factors such as weather, traffic, and public safety events, thereby realizing all-round and multi-angle monitoring and analysis of personal safety risks. This way of multi-source data fusion greatly enriches the dimensions of risk assessment, making the risk assessment results more comprehensive and accurate. By comprehensively considering internal and external factors, the embodiment of the present application can more accurately identify potential risk points and avoid the one-sidedness and limitations that may be brought by a single data source.

[0132] Furthermore, the specific implementation method of risk prediction based on the personal risk identification model will be introduced in detail below. In an exemplary embodiment, as Figure 5 shown, the implementation process of the personal risk identification model of a personal risk identification method based on multi-source data information is provided. In this embodiment, the implementation process includes the following steps:

[0133] Step 501, Data Feature Engineering: According to the requirements of power supply enterprise employees' personal safety risk assessment, further feature engineering is carried out on the processed data, including feature construction, feature scaling, and encoding conversion, to enhance the interpretability of the model and improve the learning efficiency of the model. Specifically, categorical variables are converted into numerical features through one-hot encoding and label encoding, continuous features are normalized, and composite features are constructed according to business logic.

[0134] Step 502, Algorithm Selection and Model Construction: The server constructs the matching personal risk identification models to be verified through logistic regression, decision tree, random forest, gradient boosting tree, and neural network algorithms respectively, and performs cross-validation on each personal risk identification model to be verified to obtain the model performance of each personal risk identification model to be verified. At this time, the server needs to select logistic regression, decision tree, random forest, gradient boosting decision tree (GBDT), and neural network algorithms to construct personal risk identification models respectively based on the complexity and diversity of multi-source data information, as well as the real-time and accuracy requirements of personal risk identification; in the model construction stage, cross-validation is first performed to evaluate the model performance under different algorithms and parameter combinations, select the optimal model configuration, and further optimize the prediction accuracy and generalization ability of the model by adjusting the hyperparameters of the model (depth of the tree, learning rate, regularization coefficient).

[0135] Step 503, Model Training and Optimization: After determining the initial personal risk identification model, the server uses the preprocessed and feature-engineered dataset to train the model. During the training process, a suitable loss function (either cross-entropy loss or mean squared error) is used to measure the difference between the model prediction value and the actual value, and the model parameters are updated through the gradient descent optimization algorithm to minimize the loss function.

[0136] Step 504, Model Evaluation and Interpretation: After adjusting the model parameters of the initial personal risk identification model, an independent test set is used to evaluate the model to ensure that the model has good generalization ability, that is, it can maintain stable prediction performance on unknown data. When evaluating the model, not only common indicators such as accuracy, recall rate, and F1 score are concerned, but also the interpretability, stability, and real-time performance of the model are combined. Specifically, a model with strong interpretability helps the enterprise understand the internal mechanism of risk factors and determine targeted prevention and control measures; a model with good stability can maintain consistent prediction performance at different time points and data distributions; real-time performance requires the model to be able to complete the prediction task in a short time to meet the enterprise's need to quickly respond to risk changes.

[0137] Step 505, Model Deployment and Monitoring: Deploy the trained and evaluated personal risk identification model into the actual business system, integrate it with the production management system, human resource management system, and safety management system of the power supply enterprise to achieve real-time monitoring and analysis. During the operation of the model, by setting up a monitoring mechanism, regularly check the prediction accuracy and stability of the model, promptly discover and handle model degradation problems. At the same time, according to business requirements and data changes, regularly update and optimize the model to maintain its optimal performance.

[0138] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are shown in sequence according to the arrows, these steps do not necessarily have to be executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages do not necessarily have to be executed at the same time, but can be executed at different times. The execution order of these steps or stages does not necessarily have to be sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0139] Based on the same inventive concept, the embodiments of the present application also provide a personal risk identification device for implementing the method for personal risk identification involved above. The solution provided by this device to solve the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the personal risk identification device provided below can refer to the limitations on the method for personal risk identification in the above text, and will not be repeated here.

[0140] In an exemplary embodiment, as Figure 6 shown, a personal risk identification device is provided, including: a data information collection module 602, a data information processing module 604, a data analysis module 606, and a model prediction module 608, where:

[0141] The data information collection module 602 is used to collect multiple personal safety data information from a multi-source information system, and each personal safety data information is respectively matched with corresponding risk information; the risk information includes a risk level and a risk type;

[0142] The data information processing module 604 is used to preprocess the multiple personal safety data information to obtain multiple preprocessed data information; the preprocessed data information conforms to the data quality policy;

[0143] A data analysis module 606, configured to perform data analysis on each piece of preprocessed data information based on data mining techniques, and extract personal safety features from each piece of preprocessed data information;

[0144] A model prediction module 608, configured to perform risk prediction through the constructed initial personal risk identification model based on each personal safety feature, and obtain prediction risk information corresponding to each piece of personal safety data information.

[0145] In an exemplary embodiment, the personal risk identification device further includes a model parameter tuning module, a risk prediction module, and a warning and intervention module;

[0146] The model parameter tuning module is configured to adjust the model parameters of the initial personal risk identification model through each prediction risk information and each risk information, and obtain a personal risk identification model for personal risk identification;

[0147] The risk prediction module is configured to deploy the personal risk identification model to the actual business system, and predict the real-time risk information of the power supply enterprise employees based on the real-time personal safety data information of the power supply enterprise employees through the personal risk identification model; the real-time risk information includes a real-time risk level and a real-time risk type; the real-time personal safety data information includes at least a working environment and a behavior pattern;

[0148] The warning and intervention module is configured to determine a risk warning matching the real-time risk information, and determine an intervention measure matching the real-time risk information.

[0149] In an exemplary embodiment, the warning and intervention module is specifically configured to determine the potential risks of the power supply enterprise employees based on the real-time risk level and the real-time risk type; trigger a risk warning when the potential risks of the power supply enterprise employees reach a preset threshold.

[0150] In an exemplary embodiment, the warning and intervention module is specifically configured to determine intervention strategies corresponding to each real-time risk type; establish a feedback opinion collection mechanism according to the real-time risk type in the real-time risk information, and collect feedback opinions for different real-time risk types respectively based on the feedback opinion collection mechanism; optimize and adjust the intervention strategies of the real-time risk type through the feedback opinions of each real-time risk type.

[0151] In an exemplary embodiment, the personal risk identification device further includes a model construction module;

[0152] A model construction module is configured to respectively construct a to-be-verified personal risk identification model that matches based on various personal safety features through multiple model construction algorithms; the multiple model construction algorithms include at least one of a logistic regression algorithm, a decision tree algorithm, a random forest algorithm, a gradient boosting tree algorithm, and a neural network algorithm; based on various personal safety features, respectively perform cross-validation on each to-be-verified personal risk identification model to obtain the model performance corresponding to each to-be-verified personal risk identification model; determine the to-be-verified personal risk identification model with the optimal model performance as the initial personal risk identification model.

[0153] In an exemplary embodiment, a data analysis module is configured to perform data analysis on each preprocessed data information by using any one of a correlation analysis, a principal component analysis, and a mutual information method to obtain a data analysis result; based on the data analysis result, screen out the feature subset with the highest relevance and the largest amount of information related to personal safety from each preprocessed data information, and determine the feature subset as the personal safety feature;

[0154] A model prediction module is configured to identify potential patterns and association relationships in each personal safety feature by using at least one of a clustering analysis and an association rule mining technique through the constructed initial personal risk identification model; adopt at least one of a time series analysis method and a machine learning regression model method through the initial personal risk identification model, and predict the predicted risk information corresponding to each personal safety data information according to the collected potential patterns and association relationships.

[0155] Each module in the above personal risk identification device can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in the processor of the computer device in hardware form or be independent of it, or can be stored in the memory of the computer device in software form, so as to facilitate the processor to call and execute the operations corresponding to the above respective modules.

[0156] In an exemplary embodiment, a computer device is provided. The computer device can be a server, and its internal structure diagram can be as Figure 7As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data related to the embodiments of the present application, such as personal safety data information and risk information. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals through a network connection. When the computer program is executed by the processor, it implements a method for identifying personal risks.

[0157] Those skilled in the art can understand that Figure 7 the structure shown in is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0158] In one embodiment, a computer device is further provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps in the above method embodiments are implemented.

[0159] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0160] In one embodiment, a computer program product is provided, including a computer program. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0161] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0162] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in this application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in this application can be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, data processing logics based on quantum computing, artificial intelligence (AI) processors, etc., without limitation.

[0163] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as within the scope recorded in this application.

[0164] The above-described embodiments merely represent several implementation manners of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A method for identifying personal risk, characterized in that: The method comprises: Collect multiple personal safety data information from a multi-source information system, each of which is matched with corresponding risk information; the risk information includes risk level and risk type; the multi-source information system includes an external system and various business systems under the power supply enterprise, and each of which includes at least workload and stress assessment data, personal health monitoring data, and external threat and risk intelligence data; Preprocessing the plurality of personal safety data information to obtain a plurality of preprocessed data information; the preprocessed data information complies with a data quality strategy; the data quality strategy at least includes: a unified format, no duplication of data, no missing data values, and no abnormal data values; Based on data mining technology, data analysis is performed on each of the preprocessed data information to extract personal safety features from each of the preprocessed data information; the personal safety features are obtained by constructing features, scaling features, and encoding the preprocessed data information according to the needs of personal safety risk assessment of power supply enterprise employees through data feature engineering; Based on each of the personal safety features, risk prediction is performed through the constructed initial personal risk identification model to obtain predicted risk information corresponding to each of the personal safety data information; Adjusting the model parameters of the initial personal risk identification model through each of the predicted risk information and each of the risk information, and evaluating the initial personal risk identification model after the model parameter adjustment using an independent test set to obtain a personal risk identification model for personal risk identification; The personal risk identification model is deployed into the actual business system, integrated with the production management system, human resource management system, and safety management system of the power supply enterprise, and the real-time risk information of the power supply enterprise employees is predicted based on the real-time personal safety data information of the power supply enterprise employees through the personal risk identification model; the real-time risk information includes the real-time risk level and the real-time risk type; the real-time personal safety data information includes at least the working environment and behavior pattern; the real-time risk type includes at least the risk of electric shock, the risk of falling from a height, the risk of mechanical injury, the risk of fire and explosion, the risk of chemical leakage, and the risk of radiation injury; Determine the risk warning that matches the real-time risk level and the real-time risk type, and determine the intervention strategy corresponding to each of the real-time risk types; establish a feedback collection mechanism according to the real-time risk type in the real-time risk information, and collect feedback for different real-time risk types based on the feedback collection mechanism; optimize and adjust the intervention strategy of the real-time risk type through the feedback of each real-time risk type; wherein, when the real-time risk type is the electric shock risk, the matched intervention measure is to strengthen the on-site safety inspection; when the real-time risk type is the high-altitude fall risk, the matched intervention measure is to strengthen the use of safety belts and the setting of protective nets; when the real-time risk type is the mechanical injury risk, the matched intervention measure is to regularly maintain and inspect mechanical equipment; when the real-time risk type is the fire and explosion risk, the matched intervention measure is to strengthen the construction and maintenance of fire-fighting facilities; when the real-time risk type is the chemical leakage risk, the matched intervention measure is to strictly manage the storage and use areas of chemicals; when the real-time risk type is the radiation injury risk, the matched intervention measure is to strictly control the radioactive source; During the operation of the personal risk identification model, the personal risk identification model is regularly checked through a monitoring mechanism to discover and handle model degradation problems.

2. The method according to claim 1, characterized in that The determining of a risk warning matching the real-time risk information includes: Determining the potential risk of the power supply enterprise employees based on the real-time risk level and the real-time risk type; When the potential risk of the employees of the power supply enterprise reaches a preset threshold, a risk warning is triggered.

3. The method according to claim 1 or 2, characterized in that: The initial personal risk identification model is constructed as follows: Based on each of the personal safety features, a matched personal risk identification model to be verified is constructed respectively by using a plurality of model construction algorithms; the plurality of model construction algorithms include at least one of a logistic regression algorithm, a decision tree algorithm, a random forest algorithm, a gradient boosting tree algorithm, and a neural network algorithm; Based on the personal safety features, cross-validate the personal risk identification models to be verified to obtain the model performances corresponding to the personal risk identification models to be verified; The personal risk identification model to be verified with the best model performance is determined as the initial personal risk identification model.

4. The method according to claim 1 or 2, characterized in that: Based on each of the personal safety features, risk prediction is performed through the constructed initial personal risk identification model to obtain predicted risk information corresponding to each of the personal safety data information, including: Using the constructed initial personal risk identification model, at least one of cluster analysis and association rule mining techniques is used to identify potential patterns and association relationships in each of the personal safety features; The initial personal risk identification model adopts at least one of a time series analysis method and a machine learning regression model method to predict the predicted risk information corresponding to each of the personal safety data information based on the collected potential patterns and the association relationships.

5. A personal risk identification device, characterized in that: The device comprises: A data information collection module is used to collect multiple personal safety data information from a multi-source information system, each of which is matched with corresponding risk information; the risk information includes risk level and risk type; the multi-source information system includes an external system and various business systems under the power supply enterprise, and each of which includes at least workload and stress assessment data, personal health monitoring data, and external threat and risk intelligence data; A data information processing module, used to pre-process the plurality of personal safety data information to obtain a plurality of pre-processed data information; the pre-processed data information complies with a data quality strategy; the data quality strategy at least includes: a unified format, no duplication of data, no missing data values, and no abnormal data values; A data analysis module, used to perform data analysis on each of the pre-processed data information based on data mining technology, and extract personal safety features from each of the pre-processed data information; A model prediction module is used to perform risk prediction based on each of the personal safety features through the constructed initial personal risk identification model to obtain predicted risk information corresponding to each of the personal safety data information; A model parameter adjustment module, used to adjust the model parameters of the initial personal risk identification model through each of the predicted risk information and each of the risk information, so as to obtain a personal risk identification model for personal risk identification; A risk prediction module is used to deploy the personal risk identification model into an actual business system, and predict the real-time risk information of the power supply enterprise employees based on the real-time personal safety data information of the power supply enterprise employees through the personal risk identification model; the real-time risk information includes the real-time risk level and the real-time risk type; the real-time personal safety data information includes at least the working environment and behavior pattern; the real-time risk type includes at least the risk of electric shock, the risk of falling from a height, the risk of mechanical injury, the risk of fire and explosion, the risk of chemical leakage, and the risk of radiation injury; An early warning and intervention module is used to determine a risk early warning that matches the real-time risk level and the real-time risk type, and to determine intervention measures that match the real-time risk type; wherein, when the real-time risk type is the electric shock risk, the matched intervention measures are to strengthen on-site safety inspections; when the real-time risk type is the height-falling risk, the matched intervention measures are to strengthen the use of safety belts and the setting of protective nets; when the real-time risk type is the mechanical injury risk, the matched intervention measures are to regularly maintain and inspect mechanical equipment; when the real-time risk type is the fire and explosion risk, the matched intervention measures are to strengthen the construction and maintenance of fire-fighting facilities; when the real-time risk type is the chemical leakage risk, the matched intervention measures are to strictly manage the storage and use areas of chemicals; when the real-time risk type is the radiation injury risk, the matched intervention measures are to strictly control the radiation source.

6. The device according to claim 5, characterized in that The early warning and intervention module is specifically used to determine the potential risks of the employees of the power supply enterprise based on the real-time risk level and the real-time risk type; and trigger a risk early warning when the potential risk of the employees of the power supply enterprise reaches a preset threshold.

7. The device according to claim 5, characterized in that The personal risk identification device also includes a model building module; The model building module is used to build the matched personal risk identification model to be verified respectively based on each of the personal safety features through multiple model building algorithms; The multiple model building algorithms include at least one of a logistic regression algorithm, a decision tree algorithm, a random forest algorithm, a gradient boosting tree algorithm, and a neural network algorithm; based on the personal safety features, each of the personal risk identification models to be verified is cross-validated to obtain the model performance corresponding to each of the personal risk identification models to be verified; the personal risk identification model to be verified with the best model performance is determined as the initial personal risk identification model.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 4 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images