Occupational disease pathogenic factor analysis method, system and equipment and computer readable storage medium

Abstract

The invention provides an occupational disease pathogenic factor analysis method, system and equipment, and the method comprises the steps: elastically collecting multi-modal exposure data, combining a dynamic weight distribution algorithm to generate a standardized feature matrix, extracting a time sequence causal path through a dynamic Bayesian network, and calculating an average causal effect. And quantifying the pathogenic factor influence degree based on the exposure intensity and the duration. According to the method, the technical problems of risk assessment deviation caused by data missing, scheme optimization lag caused by neglect of a time sequence effect and a cooperative path by a static causal model and dependence on artificial experience in a traditional method can be solved, and the accuracy and real-time performance of occupational disease pathogenic factor analysis are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of occupational disease prevention and treatment, and in particular to an occupational disease pathogenic factor analysis method, system, device and computer-readable storage medium. Background Art

[0002] Traditional occupational disease causative factor analysis methods can usually only obtain preset exposure parameters of the target occupational scenario during data collection and processing. Once the occupational scenario is changed or key data is missing, the system can neither calculate the equivalent exposure dose through environmental parameters nor dynamically switch to alternative data sources, resulting in deviations in risk assessment. In addition, existing causal analysis models are mostly static structures that do not take into account the dynamic changes and timeliness of data quality, and ignore indirect paths and synergistic effects as treatment factors, resulting in one-sided risk assessment. Therefore, there is an urgent need for a solution that can integrate flexible data collection, dynamic causal modeling, and is suitable for different occupational scenarios. Summary of the Invention

[0003] In response to the shortcomings of the existing technology, the present invention proposes a method, system, device and computer-readable storage medium for analyzing the causative factors of occupational diseases to solve the technical problems existing in the existing technology, such as large deviations in occupational disease risk assessment and inefficient improvement measures.

[0004] The technical solution adopted by the present invention is a method for analyzing the pathogenic factors of occupational diseases.

[0005] In the first possible implementation, it includes: elastically collecting multimodal exposure data of target occupational scenarios; Processing the multimodal exposure data using a dynamic weight allocation algorithm to generate a standardized feature matrix; Dynamic Bayesian network analysis is used to analyze the standardized feature matrix and extract a causal feature dataset of the impact of exposure factors on the disease; Calculating the influence of occupational disease causative factors based on the causal characteristic data set; A table of causative factor possibilities is generated based on the degree of influence.

[0006] Furthermore, multimodal exposure data for target occupational scenarios can be collected flexibly in real time, including: collecting mandatory data items, including physical exposure parameters and biomarker data; Collect optional data items, including chemical exposure data and environmental parameters; When the chemical exposure data is missing, the exposure substitute value is calculated based on the ventilation volume and contact time of the target occupational scenario using the following exposure substitute value calculation formula to replace the chemical exposure data: in, represents the exposure substitution value, V represents the ventilation volume, and t represents the contact time. represents the correction factor of the i-th environmental parameter, where is the temperature correction factor, is the humidity correction factor, is the wind speed correction factor.

[0007] Furthermore, the multimodal exposure data is processed by a dynamic weight allocation algorithm to generate a standardized feature matrix, including: obtaining the occupational scenario type and time decay coefficient of the multimodal exposure data; The dynamic weight vector is calculated according to the occupational scenario type and the time decay coefficient, and the dynamic weight vector matrix of each multimodal exposure data is allocated: in, represents the dynamic weight vector matrix of the multimodal exposure data, n represents a total of n types of data, Represents the dynamic weight vector of the nth class of data; multiplying the dynamic weight vector matrix and the multimodal exposure data element-wise to obtain a weighted data matrix; The weighted data matrix is ​​normalized by sub-modality, and each modality data is aligned according to the time window to generate a standardized feature matrix.

[0008] Furthermore, a dynamic weight vector is calculated according to the occupational scenario type and the time decay coefficient. The dynamic weight vector of the i-th category data is calculated according to the occupational scenario type and the time decay coefficient using the method shown in the following dynamic weight vector calculation formula: in, Represents the dynamic weight vector of the i-th category data, represents the scene type coefficient, represents the signal-to-noise ratio of the i-th category data, represents the decay rate constant, Indicates the delay between the current time and the data collection time.

[0009] Furthermore, a dynamic Bayesian network is used to analyze the standardized feature matrix to extract a causal feature data set of the impact of exposure factors on the disease, including: dividing the standardized feature matrix into multiple sub-matrices according to a preset time window, and mapping each sub-matrix to a node set of a dynamic Bayesian network, wherein the node set includes exposure nodes, intermediate biomarker nodes, and disease endpoint nodes; Based on the temporal dependencies in the standardized feature matrix, adding time-delayed directed edges between nodes and learning the network structure through conditional independence tests; Learning a conditional probability table from the standardized feature matrix to generate a dynamic Bayesian network model; Extracting the causal path from the exposure factor to the disease endpoint from the dynamic Bayesian network model and calculating the path transmission probability; Conduct intervention simulation on the exposure factors in the standardized characteristic matrix, setting high exposure level and low exposure level scenarios respectively; The difference in the probability of disease endpoints occurring under high-exposure and low-exposure scenarios is calculated through the forward propagation of the dynamic Bayesian network. The average causal effect is calculated based on the difference in the probability of occurrence of the disease endpoint under the high exposure level scenario and the low exposure level scenario using the method shown in the target average causal effect calculation formula below: in, represents the average causal effect of the ith exposure factor, represents the probability of occurrence of high exposure levels, Indicates the probability of occurrence of low exposure levels; The causal paths whose average causal effect value exceeds the preset threshold are screened to generate a causal feature dataset.

[0010] Furthermore, the influence of occupational disease causative factors is calculated based on the causal characteristic data, including: obtaining exposure intensity and cumulative exposure duration of occupational disease hazard factors, wherein the exposure intensity is used to represent the multiple of the worker's exposure to occupational disease hazard factors per unit time compared with the safety limit; The impact score is calculated based on the exposure intensity and duration using the method shown in the impact score calculation formula below: in, Indicates the degree of influence, represents the average causal effect of the ith exposure factor, Indicates the exposure intensity, Indicates duration.

[0011] Furthermore, generating causative factors according to the impact degree includes: arranging the exposure factors in descending order according to the impact degree scores, and generating a causative factor matching list.

[0012] In combination with the first feasible method, in the second feasible method, a system for implementing occupational disease causative factor analysis and improvement method is included, including: a data acquisition module, used to flexibly collect multimodal exposure data of target occupational scenarios; a data processing module, used to process the multimodal exposure data through a dynamic weight distribution algorithm to generate a standardized feature matrix; an exposure factor extraction module, used to analyze the standardized feature matrix using a dynamic Bayesian network to extract a causal feature data set of the impact of exposure factors on the disease; an influence calculation module, used to calculate the influence of occupational disease causative factors based on the causal feature data set; and an output module, used to output a causal factor possibility table generated according to the influence.

[0013] In combination with the first possible implementation method, in a third possible implementation method, a device is included, including a memory and a processor, the memory stores computer instructions that can be run on the processor, and the processor executes a method for analyzing occupational disease causative factors when running the computer instructions.

[0014] In combination with the first possible implementation, a fourth possible implementation includes a computer-readable storage medium having computer instructions stored thereon. When the computer instructions are executed, a method for analyzing the causative factors of occupational diseases can be implemented.

[0015] It can be seen from the above technical solution that the beneficial technical effects of the present invention are as follows: 1. Through the flexible data collection mechanism of mandatory / optional / alternative data, alternative data is generated when chemical exposure data is missing, ensuring analysis continuity, solving the assessment interruption problem caused by missing data, and reducing the risk assessment deviation rate.

[0016] 2. The dynamic weight allocation algorithm combines data quality, timeliness and industry characteristics to generate a standardized feature matrix, which is then analyzed through a dynamic Bayesian network to improve the accuracy of risk assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly describes the drawings required for the specific embodiments or the description of the prior art. Similar elements or parts are generally identified by similar reference numerals throughout the drawings. Elements or parts in the drawings are not necessarily drawn to scale.

[0018] Figure 1 This is a flow chart of Example 1 of the present invention; Figure 2 A flow chart for generating a standardized feature matrix according to embodiment 1 of the present invention; Figure 3 This is a system structure diagram of Example 2 of the present invention. DETAILED DESCRIPTION

[0019] The following embodiments of the technical solution of the present invention will be described in detail with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention and are therefore only examples and are not intended to limit the scope of protection of the present invention.

[0020] It should be noted that, unless otherwise specified, the technical or scientific terms used in this application should have the common meanings understood by those skilled in the art to which the present invention belongs.

[0021] Example 1 This embodiment provides a method for analyzing the causative factors of occupational diseases. The working principle of embodiment 1 is described in detail below: The flow chart of this embodiment is as follows Figure 1 As shown, it includes: flexibly collecting multimodal exposure data of target occupational scenarios; normalizing the multimodal exposure data through a dynamic weight allocation algorithm to generate a standardized feature matrix; using a dynamic Bayesian network to analyze the standardized feature matrix and extract a causal feature data set of the impact of exposure factors on diseases; calculating the influence of occupational disease causal factors based on the causal feature data set; generating an occupational disease improvement plan based on the influence; collecting feedback data from users on their execution of the occupational disease improvement plan; and dynamically adjusting the occupational disease improvement plan through a reinforcement learning model based on the feedback data to generate an optimized occupational disease improvement plan.

[0022] In this embodiment, multimodal exposure data of target occupational scenarios is collected in real time and flexibly, including: Mandatory data items include physical exposure parameters (noise decibel value, vibration acceleration) and biomarker data (specific antibody concentration); optional data items include chemical exposure data (occupational exposure limit) and environmental parameters (CO2 concentration).

[0023] When the chemical exposure data is missing, the exposure substitute value is calculated based on the ventilation volume and contact time of the target occupational scenario using the following exposure substitute value calculation formula to replace the chemical exposure data: in, represents the exposure substitution value, V represents the ventilation volume, and t represents the contact time. represents the correction factor of the i-th environmental parameter, where is the temperature correction factor, is the humidity correction factor, is the wind speed correction factor.

[0024] By introducing correction factors for environmental parameters, in occupational fields where chemical exposure values ​​cannot be monitored or when chemical exposure value data are missing, the correction factors can be used to replace the chemical exposure data, thereby quantifying the impact of environmental parameters in different industries. This improvement improves the versatility and accuracy of occupational disease risk assessment in data-missing scenarios.

[0025] The multimodal exposure data are normalized by a dynamic weight allocation algorithm to generate a standardized feature matrix such as Figure 2 Shown, including: Obtain the occupational scenario type and time attenuation coefficient of the multimodal exposure data and assign a dynamic weight vector , where n represents a total of n types of data; in this embodiment, the ratio of signal power to noise power is calculated using the original sensor signal; and the scene type is determined based on the preset label input by the user or the clustering result of the environmental parameters.

[0026] In the dynamic weight vector, the dynamic weight vector of the i-th type of data satisfies the formula: in, Represents the dynamic weight vector of the i-th category data, represents the scene type coefficient, represents the signal-to-noise ratio of the i-th category data, represents the decay rate constant, Indicates the delay between the current time and the data collection time, so is the time decay term, which is used to reflect the timeliness of data. =0 means real-time data, the attenuation term is 1, and the weight is maximized; Increase means historical data, weight decays exponentially, and the decay rate constant Controls the decay rate. This formula is used to quantify the contribution of data quality to the weight and is calculated through the scene type coefficient To balance the priority of signal-to-noise ratio and time weight, set a larger scene type coefficient when applicable to noise-sensitive parameter scenes Used to emphasize data quality; when applicable to noise-sensitive parameter scenarios, set a smaller scenario type coefficient Used to weaken the impact of transient noise.

[0027] multiplying the dynamic weight vector and the multimodal exposure data element-wise to obtain a weighted data matrix; The weighted data matrix is ​​subjected to modal normalization. This includes performing a logarithmic transformation on the weighted chemical concentration data to eliminate dimensional differences in concentrations; performing Z-score normalization on physical exposures; performing threshold normalization on biological factors; and performing piecewise discretization on environmental parameters according to relevant international standards, such as humidity classification according to ISO 7730. The modal data are then aligned by time window to generate a standardized feature matrix. These modal normalization steps can all be implemented using existing technologies and are therefore not further detailed in this application.

[0028] A dynamic Bayesian network is used to analyze the standardized feature matrix to extract a causal feature data set of the impact of exposure factors on the disease, including: dividing the standardized feature matrix into multiple sub-matrices according to a preset time window, each sub-matrix is ​​mapped to a node set of a dynamic Bayesian network, and the node set includes exposure nodes, intermediate biomarker nodes and disease endpoint nodes; based on the temporal dependency in the standardized feature matrix, time-delayed directed edges are added between nodes, and the network structure is learned through a conditional independence test; a conditional probability table is learned from the standardized feature matrix to generate a dynamic Bayesian network model; the causal path from the exposure factor to the disease endpoint is extracted from the dynamic Bayesian network model, and the path transmission probability is calculated; an intervention simulation is performed on the exposure factors in the standardized feature matrix, and high exposure level and low exposure level scenarios are set respectively; the difference in the probability of occurrence of the disease endpoint under the high exposure level scenario and the low exposure level scenario is calculated through the forward propagation of the dynamic Bayesian network; the average causal effect is calculated based on the difference in the probability of occurrence of the disease endpoint under the high exposure level scenario and the low exposure level scenario using the method shown in the following target average causal effect calculation formula: in, represents the average causal effect of the ith exposure factor, represents the probability of occurrence of high exposure levels, Indicates the probability of occurrence of low exposure levels; The causal paths whose average causal effect value exceeds the preset threshold are screened to generate a causal feature dataset.

[0029] The influence of occupational disease causative factors is calculated based on the causal characteristic data, including: Obtain the exposure intensity and cumulative duration of exposure to occupational disease hazard factors. The exposure intensity is used to indicate the multiple of the worker's exposure to occupational disease hazard factors exceeding the safety limit per unit time. Calculate the impact score based on the exposure intensity and duration using the following impact score calculation formula: in, Indicates the degree of influence, represents the average causal effect of the ith exposure factor, Indicates the exposure intensity, Indicates duration.

[0030] Generating pathogenic factors according to the impact degree includes: arranging the exposure factors in descending order according to the impact degree scores, and generating a pathogenic factor matching list.

[0031] After obtaining the impact, it also includes generating an occupational disease improvement plan based on the impact, including: arranging the scores in descending order according to the impact, and generating a priority list of interventions for pathogenic factors; screening candidate plans for occupational disease improvement plans from a preset intervention measures database; evaluating the effectiveness and pertinence of the candidate plans through a reinforcement learning model combined with the individual biomarker data and exposure scenario characteristics of the target user, and selecting the optimal health intervention plan as the final occupational disease improvement plan.

[0032] The intervention database contains: A database of engineering control measures, including physical isolation measures used to reduce occupational disease exposure levels; a database of management control measures, including binding management systems used to reduce the cumulative effects of occupational disease hazards; and a database of personal protection measures, including personal protective equipment and dietary training programs used to block occupational disease hazards.

[0033] Collecting user feedback data based on the implementation of the occupational disease improvement plan, including: collecting the user's physiological data, behavioral data, environmental data and psychological data, wherein: the physiological data includes heart rate, blood pressure and blood oxygen saturation, which are continuously collected through wearable devices; the behavioral data includes the user's completion rate and execution time deviation of the occupational disease improvement plan, which are collected through mobile terminal log records and computer vision behavior recognition systems; the environmental data includes indoor temperature, humidity, air quality and noise level, which are monitored in real time through the Internet of Things sensor network; the psychological data includes emotional state vector and stress index, which are analyzed and output through a speech sentiment analysis model.

[0034] Example 2 This embodiment provides a system for analyzing the causes of occupational diseases and improving methods, such as Figure 3 As shown, it includes: a data acquisition module for flexibly collecting multimodal exposure data of target occupational scenarios; A data processing module is used to process the multimodal exposure data through a dynamic weight allocation algorithm to generate a standardized feature matrix; an exposure factor extraction module is used to analyze the standardized feature matrix using a dynamic Bayesian network to extract a causal feature data set of the impact of exposure factors on the disease; an influence degree calculation module is used to calculate the influence of occupational disease causal factors based on the causal feature data set; and an output module is used to output a causal factor possibility table generated based on the influence degree.

[0035] Furthermore, in this embodiment, after generating the pathogenic factor possibility table, the influence calculation module further includes: Based on the causative factor possibility table, candidate occupational disease improvement plans are screened from a preset intervention measures database.

[0036] The intervention measures database in this embodiment includes: A database of engineering control measures, including physical isolation measures, used to reduce occupational disease exposure levels; A database of management control measures, including management systems used to reduce the cumulative effects of occupational hazards; A database of personal protective measures, including personal protective equipment and dietary training programs used to block occupational hazards.

[0037] Collecting user feedback data based on the occupational disease improvement plan, including collecting user physiological data, behavioral data, environmental data, and psychological data, including: The physiological data includes heart rate, blood pressure and blood oxygen saturation, which are continuously collected by wearable devices; The behavioral data includes the user's completion rate and execution time deviation of the occupational disease improvement plan, which is collected through mobile terminal log records and computer vision behavior recognition system; The environmental data includes indoor temperature, humidity, air quality and noise level, which are monitored in real time through an IoT sensor network; The psychological data includes an emotional state vector and a stress index, which are analyzed and output by a speech emotion analysis model.

[0038] Combined with the feedback data, the effectiveness and pertinence of the candidate solutions are evaluated through a reinforcement learning model to generate an optimized occupational disease improvement plan.

[0039] Example 3 This embodiment provides a device including a memory and a processor, wherein the memory stores computer instructions that can be executed on the processor, and the processor executes the method for analyzing the causative factors of occupational diseases when executing the computer instructions.

[0040] Example 4 This embodiment provides a computer-readable storage medium having computer instructions stored thereon. When the computer instructions are executed, a method for analyzing the causative factors of occupational diseases can be implemented.

[0041] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some or all of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and description of the present invention.

Claims

1. A method for analyzing the pathogenic factors of occupational diseases, characterized in that: include: Flexible collection of multimodal exposure data for target occupational scenarios; Processing the multimodal exposure data using a dynamic weight allocation algorithm to generate a standardized feature matrix; Dynamic Bayesian network analysis is used to analyze the standardized feature matrix and extract a causal feature dataset of the impact of exposure factors on the disease; Calculating the influence of occupational disease causative factors based on the causal characteristic data set; A table of causative factor possibilities is generated based on the degree of influence.

2. The method for analyzing the causative factors of occupational diseases according to claim 1, characterized in that: Real-time and flexible collection of multimodal exposure data for target occupational scenarios, including: Collect mandatory data items, including physical exposure parameters and biomarker data; Collect optional data items, including chemical exposure data and environmental parameters; When the chemical exposure data is missing, the exposure substitute value is calculated based on the ventilation volume and contact time of the target occupational scenario using the following exposure substitute value calculation formula to replace the chemical exposure data: in, represents the exposure substitution value, V represents the ventilation volume, and t represents the contact time. represents the correction factor of the i-th environmental parameter, where is the temperature correction factor, is the humidity correction factor, is the wind speed correction factor.

3. The method for analyzing the causative factors of occupational diseases according to claim 2, characterized in that: The multimodal exposure data are processed by a dynamic weight allocation algorithm to generate a standardized feature matrix, including: Obtaining occupational scenario types and time decay coefficients of the multimodal exposure data; The dynamic weight vector is calculated according to the occupational scenario type and the time decay coefficient, and the dynamic weight vector matrix of each multimodal exposure data is allocated: in, represents the dynamic weight vector matrix of the multimodal exposure data, n represents a total of n types of data, Represents the dynamic weight vector of the nth class of data; multiplying the dynamic weight vector matrix and the multimodal exposure data element-wise to obtain a weighted data matrix; The weighted data matrix is ​​normalized by sub-modality, and each modality data is aligned according to the time window to generate a standardized feature matrix.

4. The method for analyzing the causative factors of occupational diseases according to claim 3, characterized in that: The dynamic weight vector is calculated according to the occupational scenario type and the time decay coefficient. The dynamic weight vector of the i-th category data is calculated according to the occupational scenario type and the time decay coefficient using the method shown in the following dynamic weight vector calculation formula: in, Represents the dynamic weight vector of the i-th category data, represents the scene type coefficient, represents the signal-to-noise ratio of the i-th type of data, represents the decay rate constant, Indicates the delay between the current time and the data collection time.

5. The method for analyzing the causative factors of occupational diseases according to claim 3, characterized in that: Dynamic Bayesian network analysis was used to analyze the standardized feature matrix and extract the causal feature dataset of the impact of exposure factors on the disease, including: Dividing the standardized feature matrix into a plurality of sub-matrices according to a preset time window, wherein each sub-matrix is ​​mapped to a node set of a dynamic Bayesian network, wherein the node set includes an exposure node, an intermediate biomarker node, and a disease endpoint node; Based on the temporal dependencies in the standardized feature matrix, adding time-delayed directed edges between nodes and learning the network structure through conditional independence tests; Learning a conditional probability table from the standardized feature matrix to generate a dynamic Bayesian network model; Extracting the causal path from the exposure factor to the disease endpoint from the dynamic Bayesian network model and calculating the path transmission probability; Conduct intervention simulation on the exposure factors in the standardized characteristic matrix, setting high exposure level and low exposure level scenarios respectively; The difference in the probability of disease endpoints occurring under high-exposure and low-exposure scenarios is calculated through the forward propagation of the dynamic Bayesian network. The average causal effect is calculated based on the difference in the probability of occurrence of the disease endpoint under the high exposure level scenario and the low exposure level scenario using the method shown in the target average causal effect calculation formula below: in, represents the average causal effect of the ith exposure factor, represents the probability of occurrence of high exposure levels, Indicates the probability of occurrence of low exposure levels; The causal paths whose average causal effect value exceeds the preset threshold are screened to generate a causal feature dataset.

6. The method for analyzing the causative factors of occupational diseases according to claim 5, characterized in that: The influence of occupational disease causative factors is calculated based on the causal characteristic data, including: Obtaining the exposure intensity and cumulative duration of exposure to occupational disease hazard factors, where the exposure intensity is used to indicate the multiple of the worker's exposure to occupational disease hazard factors exceeding the safety limit per unit time; The impact score is calculated based on the exposure intensity and duration using the method shown in the impact score calculation formula below: in, Indicates the degree of influence, represents the average causal effect of the ith exposure factor, Indicates the exposure intensity, Indicates duration.

7. The method for analyzing the causative factors of occupational diseases according to claim 6, characterized in that: Based on the impact, the causative factors are generated, including: The exposure factors are arranged in descending order according to the impact scores to generate a matching list of pathogenic factors.

8. A system for implementing the method for analyzing occupational disease causative factors according to any one of claims 1 to 7, characterized in that: include: Data collection module, used to flexibly collect multimodal exposure data of target occupational scenarios; A data processing module, configured to process the multimodal exposure data using a dynamic weight allocation algorithm to generate a standardized feature matrix; An exposure factor extraction module, configured to analyze the standardized feature matrix using a dynamic Bayesian network to extract a causal feature dataset of the effects of exposure factors on the disease; An influence calculation module, used to calculate the influence of occupational disease causative factors based on the causal feature data set; The output module is used to output a pathogenic factor possibility table generated according to the influence degree.

9. A device comprising a memory and a processor, wherein the memory stores computer instructions that can be executed on the processor, characterized in that: When the processor runs the computer instructions, it executes the method for analyzing occupational disease causative factors according to any one of claims 1 to 7.

10. A computer-readable storage medium having computer instructions stored thereon, characterized in that: When the computer instructions are executed, the method for analyzing the causative factors of occupational diseases as claimed in any one of claims 1 to 7 can be implemented.

Images