Intelligent supervision method and system based on big data

Abstract

The invention relates to the technical field of data management, in particular to an intelligent supervision method and system based on big data, and the method comprises the following steps: S1, multi-source data collection: collecting multi-source data in real time; s2, data preprocessing and spatio-temporal feature extraction: preprocessing the collected multi-source data, and extracting spatio-temporal features from the preprocessed multi-source data; s3, performing multi-dimensional risk assessment based on spatial-temporal characteristics: performing risk assessment on behaviors and states of the supervision objects by applying a multi-dimensional risk assessment algorithm; s4, dynamic early warning: in combination with a preset early warning rule, automatically triggering dynamic early warning, and providing a corresponding risk coping suggestion; and S5, supervision strategy optimization: optimizing and updating a supervision strategy through a self-learning mechanism based on a risk assessment result and feedback of dynamic early warning. According to the invention, the accuracy and response speed of risk identification are improved, so that the supervision work is more flexible and efficient.

Description

Technical Field

[0001] The present invention relates to the field of data management technology, and in particular to a smart supervision method and system based on big data. Background Art

[0002] With the rapid development of big data, the Internet of Things, and artificial intelligence technologies, regulatory systems in various industries have gradually introduced these advanced technologies to improve the level of intelligence in monitoring and management. Through real-time collection and analysis of multi-source data, regulatory systems can better cope with complex and changing environments and ensure the security and stability of business processes, equipment status, and environmental conditions. Against this background, smart regulatory systems have begun to be applied to various high-risk industries, such as manufacturing, energy management, transportation and logistics, to improve the efficiency of risk identification, early warning, and response.

[0003] Although existing smart supervision systems have made progress in multi-source data collection and monitoring, existing technologies still have several shortcomings. First, the accuracy and real-time performance of data processing and analysis are relatively weak, which makes it difficult for the system to respond quickly to potential risks in complex scenarios. Second, existing systems usually lack in-depth analysis of spatiotemporal characteristics and cannot fully utilize the time and space dimensions in the data for accurate risk prediction. In addition, the formulation of supervision strategies often relies on preset rules and lacks adaptive capabilities, making it difficult to adjust strategies in a timely manner according to the dynamically changing risk environment, resulting in limited flexibility and self-optimization capabilities of the system in long-term operation.

[0004] In view of the shortcomings of the existing technology, the present invention proposes a smart supervision method and system based on big data, which improves the risk identification accuracy and response speed of the system, enhances the adaptability and continuous optimization ability of the system in a complex and dynamic environment, and realizes intelligent and efficient risk management and decision support. Summary of the invention

[0005] The present invention provides a smart supervision method and system based on big data.

[0006] A smart supervision method based on big data includes the following steps: S1, multi-source data collection: through sensors and data collection devices deployed at multiple monitoring points, multi-source data, including environmental data, business process data and equipment status data, are collected in real time; S2, data preprocessing and spatiotemporal feature extraction: preprocess the collected multi-source data, including data denoising, missing value filling and outlier identification, and extract spatiotemporal features from the preprocessed multi-source data; S3, multi-dimensional risk assessment based on spatiotemporal features: using the extracted spatiotemporal features, applying a multi-dimensional risk assessment algorithm to conduct risk assessment on the behavior and status of the regulated object, classifying the risks according to the risk assessment results, and generating corresponding regulatory strategies for risk response, including: S31, Spatiotemporal feature analysis and modeling: By analyzing the extracted spatiotemporal features, identifying the changing trends of multi-source data in different time and space dimensions, and building a spatiotemporal model to capture the dynamic changing patterns of the behavior and status of the regulated objects; S32, Risk Assessment: Based on the spatiotemporal model, a multi-dimensional risk assessment algorithm is applied to conduct risk assessment on the behavior and status of the regulated object; S33, Risk grading and strategy generation: Based on the calculation results of the multi-dimensional risk assessment algorithm, the risks are divided into different levels, and corresponding regulatory strategies are generated according to the risk levels; S4, dynamic warning: According to the graded risk assessment results and the generated supervision strategy, combined with the preset warning rules, dynamic warnings are automatically triggered and corresponding risk response suggestions are provided; S5, Supervision strategy optimization: Based on the risk assessment results and dynamic warning feedback, the supervision strategy is optimized and updated through a self-learning mechanism.

[0007] Optionally, the multi-source data collection in S1 includes: S11, sensor deployment: deploy different types of sensors and data acquisition equipment at multiple monitoring points, including environmental sensors, process monitoring equipment, and equipment status monitoring devices; S12, real-time data monitoring: Real-time monitoring and collection of multi-source data through various sensors, including environmental data (such as temperature, humidity, air quality), business process data (such as production process operation information) and equipment status data (such as equipment operating status and fault information).

[0008] Optionally, the data preprocessing and spatiotemporal feature extraction in S2 include: S21, data denoising: denoising the collected multi-source data using Gaussian filtering algorithm; S22, missing value filling: missing values ​​in multi-source data are filled by linear interpolation; S23, outlier identification: using the Z-score algorithm anomaly detection algorithm to identify outliers; S24, spatiotemporal feature extraction: extracting spatiotemporal features from the preprocessed multi-source data, including spatial features, temporal features, spatiotemporal interaction features, motion features, and environmental state features.

[0009] Optionally, the spatiotemporal feature extraction in S24 includes: S241, Spatial feature extraction: Extract spatial features from pre-processed multi-source data through spatial distance calculation, and analyze the changes of regulatory objects in different geographical locations; S242, Time Feature Extraction: Capturing time series trends by analyzing the dynamic changes of data at different time points; S243, spatiotemporal interaction feature extraction: spatiotemporal interaction features combine the dynamic changes of space and time to analyze the interactive changes of data under different spatiotemporal conditions; S244, motion feature extraction: The motion feature describes the behavior change of the object by extracting its moving trajectory and speed; S245, environmental state feature extraction: Environmental state features are extracted by taking average values ​​of changes in environmental data (such as temperature, humidity, air quality, etc.).

[0010] Optionally, the spatiotemporal feature analysis and modeling in S31 includes: S311, dynamic change trend identification: for the change patterns of multi-source data under different time and space conditions, the dynamic change trends are identified through the K-Means spatiotemporal clustering algorithm; S312, spatiotemporal model construction: Based on the dynamic change trend identification results of spatiotemporal characteristics, a spatiotemporal regression model is constructed to capture the behavior and state change patterns of the regulated objects in different time and space dimensions.

[0011] Optionally, the risk assessment in S32 includes: S321, risk assessment based on spatiotemporal model: based on the constructed spatiotemporal regression model, calculate the state value of each dimension feature prediction ; S322, Application of multi-dimensional risk assessment algorithm: Calculate the comprehensive risk value of feature data of multiple dimensions (such as time, space, state, behavior) through weighted multi-dimensional risk algorithm , in order to assess the risk level of the regulated entity.

[0012] Optionally, the risk grading and strategy generation in S33 includes: S331, Risk classification: Based on the calculation results of the multi-dimensional risk assessment algorithm , the comprehensive risk value is compared with the preset lower limit threshold and upper threshold Make comparisons and classify risks into different levels, including low risk, medium risk, and high risk; S332, strategy generation: According to different risk levels, corresponding regulatory strategies are automatically generated. For low risk levels, only information is recorded. For medium risk levels, warnings are issued and regulatory strategies are initiated. For high risk levels, emergency response procedures are triggered to adjust regulatory intensity. Regulatory strategies include adjustment of monitoring frequency, reallocation of resources, and triggering of risk response actions.

[0013] Optionally, the dynamic warning in S4 includes: S41, Dynamic warning triggering: Based on the graded risk assessment results and the generated supervision strategy, combined with the preset warning rules, it is automatically determined whether a dynamic warning needs to be triggered. The warning rules include response thresholds and strategies for different risk levels, including: Low risk warning rules: Trigger condition: When the risk value Less than the lower threshold When triggered; Early warning mechanism: Generate low-priority early warning notifications, including background logging, low-frequency email reminders, or marking the monitoring device as a state to be observed; Medium risk warning rules: Trigger condition: When the risk value Between lower threshold and upper threshold If a risk is detected, a medium-risk warning is triggered; Early warning mechanism: When a medium-risk early warning is triggered, a medium-priority early warning notification will be generated, and a reminder will be sent to the relevant responsible person via email or instant messaging tools, and the risk will be marked as an alert status on the monitoring platform; High-risk warning rules: Trigger condition: When the risk value Above upper threshold Triggered when a serious risk is detected; Early warning mechanism: When a high-risk warning is triggered, a high-priority warning will be generated, and an emergency alert will be sent to relevant personnel through multiple channels (SMS, phone calls, emails, etc.). At the same time, a red alert will be issued on the monitoring platform, requiring immediate intervention; S42, warning rules combined with strategies: Warning rules not only consider the risk level, but also combine the generated supervision strategy to determine the warning form and intensity. According to the generated supervision strategy, the warning rules dynamically adjust the alarm frequency, notification channels and response levels, including: Low risk warning: Low risk warning is only monitored regularly. When the detected risk assessment value is lower than the lower limit threshold When the system is in operation, it only indicates the existence of potential risks; Medium risk warning: Regularly monitor the risk source for medium risk warning, and increase the density of monitoring as the risk persists or increases; High-risk warning: Real-time monitoring of high-risk warnings and continuous assessment of the status of risk sources until the risk is resolved; S43, risk response advice provision: If a dynamic warning is triggered, corresponding risk response advice will be automatically generated and provided based on the risk level and the generated regulatory strategy, including strengthening monitoring, adjusting resource scheduling, and implementing emergency measures.

[0014] Optionally, the supervision strategy optimization in S5 includes: S51, Strategy optimization trigger: Based on the risk assessment results and dynamic early warning feedback information, the effectiveness of the current regulatory strategy is analyzed in real time. When it is found that the current regulatory strategy cannot cope with the risk, the regulatory strategy optimization process will be automatically triggered; S52, application of self-learning mechanism: through the self-learning mechanism, the Q-learning algorithm is used to continuously optimize and update the supervision strategy.

[0015] A smart supervision system based on big data, used to implement the above-mentioned smart supervision method based on big data, includes the following modules: Multi-source data collection module: Through sensors and data collection devices deployed at multiple monitoring points, multi-source data is collected in real time, including environmental data, business process data, and equipment status data; Data preprocessing and spatiotemporal feature extraction module: preprocess the collected multi-source data, including data denoising, missing value filling and outlier identification, and extract spatiotemporal features from the preprocessed multi-source data; Multidimensional risk assessment module: Based on the extracted spatiotemporal features, a multidimensional risk assessment algorithm is applied to assess the behavior and status of the regulated object, risks are graded according to the assessment results, and corresponding regulatory strategies are generated; Dynamic warning module: Automatically trigger dynamic warnings based on the graded risk assessment results and generated supervision strategies, combined with preset warning rules; Regulatory strategy optimization module: Based on risk assessment results and dynamic early warning feedback, the regulatory strategy is optimized and updated through a self-learning mechanism.

[0016] Beneficial effects of the present invention: The present invention comprehensively captures the dynamic changes of the regulatory objects in multiple dimensions such as space, time, and environment through multi-source data collection, data preprocessing, and spatiotemporal feature extraction. The system monitors multi-source data such as environment, business processes, and equipment status in real time through high-precision sensors, ensuring comprehensive coverage and accurate acquisition of information, providing a solid data foundation for subsequent risk assessment and regulatory strategies.

[0017] The present invention can effectively identify the potential risks of the supervision objects through the multi-dimensional risk assessment and dynamic early warning functions of spatiotemporal characteristics, and ensure the system's rapid response capabilities to different risk levels through risk classification and automated strategy generation. The system greatly improves the accuracy and response speed of risk identification by combining multi-dimensional features, automatically triggering early warning mechanisms and providing response suggestions, making supervision work more flexible and efficient.

[0018] The present invention, through the self-learning mechanism and supervision strategy optimization function, can continuously adjust and optimize the supervision strategy according to the risk assessment results and dynamic early warning feedback, to ensure that the system maintains efficient adaptability and accuracy in a complex and changing environment. Through the continuous iterative learning of the Q-learning algorithm, the intelligence level of the system is improved, so that in the long-term operation, it has higher decision-making ability and the ability to flexibly respond to different risk scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required to be used in the embodiments or the description of the prior art are briefly introduced below.

[0020] Figure 1 A schematic diagram of a supervision method flow chart of an embodiment of the present invention; Figure 2 Schematic diagram of system function modules according to an embodiment of the present invention. DETAILED DESCRIPTION

[0021] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0022] It should be noted that the references to "one embodiment", "embodiment", "exemplary embodiments", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0023] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0024] like Figure 1 As shown, a smart supervision method based on big data includes the following steps: S1, multi-source data collection: through sensors and data collection devices deployed at multiple monitoring points, multi-source data, including environmental data, business process data and equipment status data, are collected in real time; S2, data preprocessing and spatiotemporal feature extraction: preprocess the collected multi-source data, including data denoising, missing value filling and outlier identification, and extract spatiotemporal features from the preprocessed multi-source data; S3, multi-dimensional risk assessment based on spatiotemporal features: using the extracted spatiotemporal features, applying a multi-dimensional risk assessment algorithm to conduct risk assessment on the behavior and status of the regulated object, classifying the risks according to the risk assessment results, and generating corresponding regulatory strategies for risk response, including: S31, Spatiotemporal feature analysis and modeling: By analyzing the extracted spatiotemporal features, identifying the changing trends of multi-source data in different time and space dimensions, and building a spatiotemporal model to capture the dynamic changing patterns of the behavior and status of the regulated objects; S32, Risk Assessment: Based on the spatiotemporal model, a multi-dimensional risk assessment algorithm is applied to conduct risk assessment on the behavior and status of the regulated object; S33, Risk grading and strategy generation: Based on the calculation results of the multi-dimensional risk assessment algorithm, the risks are divided into different levels, and corresponding regulatory strategies are generated according to the risk levels; S4, dynamic warning: According to the graded risk assessment results and the generated supervision strategy, combined with the preset warning rules, dynamic warnings are automatically triggered and corresponding risk response suggestions are provided; S5, Supervision Strategy Optimization: Based on the risk assessment results and dynamic warning feedback, the supervision strategy is optimized and updated through the self-learning mechanism; Through the above steps, risk assessment is carried out in combination with temporal and spatial characteristics, and the regulatory strategy is continuously optimized through a self-learning mechanism. It is possible to respond to complex and changing regulatory environments in real time, improve the accuracy of risk identification and the speed of regulatory response, and at the same time enhance the system's adaptability and continuous optimization capabilities.

[0025] Multi-source data acquisition in S1 includes: S11, sensor deployment: deploy different types of sensors and data acquisition equipment at multiple monitoring points, including environmental sensors, process monitoring equipment, and equipment status monitoring devices; S12, real-time data monitoring: real-time monitoring and collection of multi-source data through various sensors, including environmental data (such as temperature, humidity, air quality), business process data (such as production process operation information) and equipment status data (such as equipment operation status and fault information); Through the above steps, different types of sensors and data acquisition equipment are deployed at multiple supervision points. The system can monitor multi-dimensional data such as environment, business processes and equipment status in real time to ensure comprehensive coverage and high-precision collection of information.

[0026] Data preprocessing and spatiotemporal feature extraction in S2 include: S21, data denoising: Use Gaussian filtering algorithm to denoise the collected multi-source data, expressed as: ; in, is the denoised data value, is the original data value, is the data center point, is the standard deviation of the Gaussian distribution; S22, missing value filling: For missing values ​​in multi-source data, linear interpolation is used to fill them, expressed as: ; in, are known data points, is the corresponding known data value, is the estimated value for the missing data points; S23, outlier identification: using the Z-score algorithm anomaly detection algorithm to identify outliers; S24, spatiotemporal feature extraction: extracting spatiotemporal features from the preprocessed multi-source data, including spatial features, temporal features, spatiotemporal interaction features, motion features, and environmental state features; Through the above content, data quality can be effectively improved, the integrity and consistency of multi-source data can be ensured, and the dynamic changes of regulatory objects can be deeply analyzed from multiple dimensions such as space, time, space-time interaction, movement and environmental status, etc., providing rich and accurate basic data for subsequent multi-dimensional risk assessment, greatly enhancing the flexibility and adaptability of the system in complex scenarios, and improving the accuracy and response speed of risk identification.

[0027] The spatiotemporal feature extraction in S24 includes: S241, spatial feature extraction: spatial features are extracted from preprocessed multi-source data through spatial distance calculation, and the changes of regulatory objects in different geographical locations are analyzed, which is expressed as: ; in, and are the spatial coordinates of the two monitoring points, is the spatial distance between two points; S242, time feature extraction: Capture time series trends by analyzing the dynamic changes of data at different time points, expressed as: ; in, and It's a different time point. Indicates time gap; S243, spatiotemporal interaction feature extraction: Spatiotemporal interaction features combine the dynamic changes of space and time to analyze the interactive changes of data under different spatiotemporal conditions, expressed as: ; S244, motion feature extraction: The motion feature describes the behavior change of the object by extracting its moving trajectory and speed, which is expressed as: ; in, is the object's speed, and are the coordinates of an object at different points in time and space; S245, environmental state feature extraction: Environmental state features are extracted by using the average value to change environmental data (such as temperature, humidity, air quality, etc.), expressed as: ; in, It is The environmental data value at a time point, is the number of time points monitored, is the average value of the environmental state; Through the above content, it is possible to comprehensively capture the dynamic changes of the regulated objects in the time and space dimensions, especially showing excellent adaptability in complex scenarios of time and space interaction, motion behavior and environmental status, providing rich feature data for multi-dimensional risk assessment, enabling the system to more accurately identify potential risks and make timely responses to abnormal situations under different time and space conditions, greatly improving the system's risk warning capabilities and regulatory efficiency.

[0028] The spatiotemporal feature analysis and modeling in S31 include: S311, dynamic change trend identification: for the change patterns of multi-source data under different time and space conditions, the dynamic change trends are identified through the K-Means spatiotemporal clustering algorithm, which is expressed as: ; in, is the number of clusters, For the clusters, is the center point of the cluster, is a data point; S312, spatiotemporal model construction: Based on the dynamic change trend identification results of spatiotemporal characteristics, a spatiotemporal regression model is constructed to capture the behavior and state change patterns of the regulated object in different time and space dimensions. The spatiotemporal regression model is expressed as: ; in, is the predicted state value, is the spatiotemporal characteristic variable, is the regression coefficient; Through the above content, it is possible to effectively identify the dynamic change trends of regulatory objects in different time and space dimensions, and build a spatiotemporal regression model to accurately predict future behaviors and states, thereby improving the system's adaptability in complex environments and capturing subtle changes in multiple dimensions. This provides more accurate data support for real-time monitoring and risk warning, while also improving the scientific nature and efficiency of regulatory decisions.

[0029] The risk assessment in S32 includes: S321, risk assessment based on spatiotemporal model: based on the constructed spatiotemporal regression model, calculate the state value of each dimension feature prediction ; S322, Application of multi-dimensional risk assessment algorithm: Calculate the comprehensive risk value of feature data of multiple dimensions (such as time, space, state, behavior) through weighted multi-dimensional risk algorithm , to assess the risk level of the regulated object, expressed as: ; in, is the comprehensive risk value, For the The weight of feature dimensions, for The state value predicted by the feature dimension, is the number of feature dimensions; Through the above content, it is possible to accurately integrate data on time, space and other characteristic dimensions to conduct comprehensive risk prediction. By using spatiotemporal regression models and weighted risk assessment algorithms, the behavior and status of the regulated objects can be dynamically analyzed.

[0030] Risk grading and strategy generation in S33 include: S331, Risk classification: Based on the calculation results of the multi-dimensional risk assessment algorithm , the comprehensive risk value is compared with the preset lower limit threshold and upper threshold For comparison, the risks are divided into different levels, including low risk, medium risk and high risk, which are expressed as: ; in, for grade; Lower threshold and upper threshold The statistical analysis method based on historical data is set, including: Data collection and collation: Collect a large amount of historical regulatory data, including past risk assessment results, incident records, and regulatory measures taken; Calculate the distribution of risk value: Statistically analyze historical risk assessment values ​​to find the comprehensive risk value The risk value is fitted using the normal distribution model, and the risk value of historical data is assumed to be According to the normal distribution, the average value is calculated and standard deviation , expressed as: ; Setting thresholds: Analyze the distribution of risk values ​​and use statistical quantiles to set thresholds and threshold , in order to classify the risk level, we choose to define the risk value below the 25th percentile as low risk, between the 25th and 75th percentile as medium risk, and above the 75th percentile as high risk. represents the 25th percentile in the risk distribution, that is, , represents the 75th percentile in the risk distribution, that is, ; Adjustment and verification: In setting and After the threshold is set, you can test and verify known events in historical data to ensure that the set threshold can reasonably divide the risk level. If the risk classification effect in the actual scenario is not ideal, you can use the new data to compare the average value. and standard deviation Re-estimate and dynamically adjust the threshold; S332, Strategy Generation: Automatically generate corresponding supervision strategies according to different risk levels. For low risk levels, only record information. For medium risk levels, issue warnings and start supervision strategies. For high risk levels, trigger emergency response procedures and adjust supervision intensity. Supervision strategies include adjustment of monitoring frequency, rescheduling of resources, and triggering of risk response actions. Through the above content, risks can be accurately divided into different levels according to actual conditions, and corresponding regulatory strategies can be generated, ensuring the scientific nature of risk assessment. Thresholds can be dynamically adjusted to adapt to changes in complex environments. Combined with an automated strategy generation mechanism, it can respond to high-risk scenarios in a timely manner and take appropriate response measures for risks of different levels, thereby improving the efficiency, accuracy and flexibility of regulatory work.

[0031] Dynamic warnings in S4 include: S41, Dynamic warning triggering: Based on the graded risk assessment results and the generated supervision strategy, combined with the preset warning rules, it is automatically determined whether a dynamic warning needs to be triggered. The warning rules include response thresholds and strategies for different risk levels, including: Low risk warning rules: Trigger condition: When the risk value Less than the lower threshold When triggered; Early warning mechanism: Generate low-priority early warning notifications, including background logging, low-frequency email reminders, or marking the monitoring device as a state to be observed; Medium risk warning rules: Trigger condition: When the risk value Between lower threshold and upper threshold If a risk is detected, a medium-risk warning is triggered; Early warning mechanism: When a medium-risk early warning is triggered, a medium-priority early warning notification will be generated, and a reminder will be sent to the relevant responsible person via email or instant messaging tools, and the risk will be marked as an alert status on the monitoring platform; High-risk warning rules: Trigger condition: When the risk value Above upper threshold Triggered when a serious risk is detected; Early warning mechanism: When a high-risk warning is triggered, a high-priority warning will be generated, and an emergency alert will be sent to relevant personnel through multiple channels (SMS, phone calls, emails, etc.). At the same time, a red alert will be issued on the monitoring platform, requiring immediate intervention; S42, warning rules combined with strategies: Warning rules not only consider the risk level, but also combine the generated supervision strategy to determine the warning form and intensity. According to the generated supervision strategy, the warning rules dynamically adjust the alarm frequency, notification channels and response levels, including: Low risk warning: Low risk warning is only monitored regularly. When the detected risk assessment value is lower than the lower limit threshold When the system is in operation, it only indicates the existence of potential risks; Medium risk warning: Regularly monitor the risk source for medium risk warning, and increase the density of monitoring as the risk persists or increases; High-risk warning: Real-time monitoring of high-risk warnings and continuous assessment of the status of risk sources until the risk is resolved; S43, risk response suggestions: If a dynamic warning is triggered, corresponding risk response suggestions are automatically generated and provided according to the risk level and the generated supervision strategy, including strengthening monitoring, adjusting resource scheduling, and implementing emergency measures to ensure that supervisors can take effective response actions based on the warning information; Through the above content, the warning frequency and notification method can be flexibly adjusted according to different risk levels, ensuring that the system only makes observation records when the risk is low, and promptly reminds relevant personnel to take preventive measures when the risk is medium. When the risk is high, high-frequency notifications through multiple channels ensure immediate response. This mechanism effectively avoids information overload caused by excessive warnings, while ensuring a rapid response to major risks, improving the flexibility and response efficiency of the system, and making supervision more targeted and accurate.

[0032] Supervision strategy optimizations in S5 include: S51, Strategy optimization trigger: Based on the risk assessment results and dynamic early warning feedback information, the effectiveness of the current regulatory strategy is analyzed in real time. When it is found that the current regulatory strategy cannot cope with the risk, the regulatory strategy optimization process will be automatically triggered; S52, application of self-learning mechanism: through the self-learning mechanism, the Q-learning algorithm is used to continuously optimize and update the supervision strategy, which is expressed as: ; in, Indicates in status Take action The expected return, is the learning rate, is the immediate return of the current feedback, is the discount factor, which is used to weight future returns, is the next state, is the next action; Through the above content, it is possible to continuously optimize and update regulatory strategies based on real-time risk assessment results and dynamic warning feedback, adapt to complex and changing environments, automatically learn and select the best strategies to improve the ability to deal with different risk scenarios. Through continuous iteration and updating, the intelligence level of the system is improved, making regulatory decisions more accurate and efficient, and enhancing the adaptability and flexibility of the system in long-term operation.

[0033] like Figure 2 As shown, a smart supervision system based on big data is used to implement the above-mentioned smart supervision method based on big data, including the following modules: Multi-source data collection module: Through sensors and data collection devices deployed at multiple monitoring points, multi-source data, including environmental data, business process data and equipment status data, can be collected in real time; Data preprocessing and spatiotemporal feature extraction module: preprocess the collected multi-source data, including data denoising, missing value filling and outlier identification, and extract spatiotemporal features from the preprocessed multi-source data; Multidimensional risk assessment module: Based on the extracted spatiotemporal features, a multidimensional risk assessment algorithm is applied to assess the behavior and status of the regulated object, risks are graded according to the assessment results, and corresponding regulatory strategies are generated; Dynamic warning module: Automatically trigger dynamic warnings based on the graded risk assessment results and generated supervision strategies, combined with preset warning rules; Regulatory strategy optimization module: Based on risk assessment results and dynamic early warning feedback, the regulatory strategy is optimized and updated through a self-learning mechanism.

[0034] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0035] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A smart supervision method based on big data, characterized in that: The following steps are involved: S1, multi-source data collection: through sensors and data collection devices deployed at multiple monitoring points, multi-source data, including environmental data, business process data and equipment status data, are collected in real time; S2, data preprocessing and spatiotemporal feature extraction: preprocess the collected multi-source data, including data denoising, missing value filling and outlier identification, and extract spatiotemporal features from the preprocessed multi-source data; S3, multi-dimensional risk assessment based on spatiotemporal features: using the extracted spatiotemporal features, applying a multi-dimensional risk assessment algorithm to conduct risk assessment on the behavior and status of the regulated object, classifying the risks according to the risk assessment results, and generating corresponding regulatory strategies for risk response, including: S31, Spatiotemporal feature analysis and modeling: By analyzing the extracted spatiotemporal features, identifying the changing trends of multi-source data in different time and space dimensions, and building a spatiotemporal model to capture the dynamic changing patterns of the behavior and status of the regulated objects; S32, Risk Assessment: Based on the spatiotemporal model, a multi-dimensional risk assessment algorithm is applied to conduct risk assessment on the behavior and status of the regulated object; S33, Risk grading and strategy generation: Based on the calculation results of the multi-dimensional risk assessment algorithm, the risks are divided into different levels, and corresponding regulatory strategies are generated according to the risk levels; S4, dynamic warning: According to the graded risk assessment results and the generated supervision strategy, combined with the preset warning rules, dynamic warnings are automatically triggered and corresponding risk response suggestions are provided; S5, Supervision strategy optimization: Based on the risk assessment results and dynamic warning feedback, the supervision strategy is optimized and updated through a self-learning mechanism.

2. According to the big data-based intelligent supervision method of claim 1, it is characterized in that: The multi-source data collection in S1 includes: S11, sensor deployment: deploy different types of sensors and data acquisition equipment at multiple monitoring points, including environmental sensors, process monitoring equipment, and equipment status monitoring devices; S12, real-time data monitoring: Real-time monitoring and collection of multi-source data through various sensors, including environmental data, business process data and equipment status data.

3. According to the big data-based intelligent supervision method of claim 1, it is characterized in that: The data preprocessing and spatiotemporal feature extraction in S2 include: S21, data denoising: denoising the collected multi-source data using Gaussian filtering algorithm; S22, missing value filling: missing values ​​in multi-source data are filled by linear interpolation; S23, outlier identification: using the Z-score algorithm anomaly detection algorithm to identify outliers; S24, spatiotemporal feature extraction: extracting spatiotemporal features from the preprocessed multi-source data, including spatial features, temporal features, spatiotemporal interaction features, motion features, and environmental state features.

4. The intelligent supervision method based on big data according to claim 3 is characterized in that: The spatiotemporal feature extraction in S24 includes: S241, Spatial feature extraction: Extract spatial features from pre-processed multi-source data through spatial distance calculation, and analyze the changes of regulatory objects in different geographical locations; S242, Time Feature Extraction: Capturing time series trends by analyzing the dynamic changes of data at different time points; S243, spatiotemporal interaction feature extraction: spatiotemporal interaction features combine the dynamic changes of space and time to analyze the interactive changes of data under different spatiotemporal conditions; S244, motion feature extraction: The motion feature describes the behavior change of the object by extracting its moving trajectory and speed; S245, environmental state feature extraction: Environmental state features are extracted by using average values ​​to measure changes in environmental data.

5. The intelligent supervision method based on big data according to claim 1 is characterized in that: The spatiotemporal feature analysis and modeling in S31 include: S311, dynamic change trend identification: for the change patterns of multi-source data under different time and space conditions, the dynamic change trends are identified through the K-Means spatiotemporal clustering algorithm; S312, spatiotemporal model construction: Based on the dynamic change trend identification results of spatiotemporal characteristics, a spatiotemporal regression model is constructed to capture the behavior and state change patterns of the regulated objects in different time and space dimensions.

6. The intelligent supervision method based on big data according to claim 5 is characterized in that: The risk assessment in S32 includes: S321, risk assessment based on spatiotemporal model: based on the constructed spatiotemporal regression model, calculate the state value of each dimension feature prediction ; S322, Application of multi-dimensional risk assessment algorithm: Calculate the comprehensive risk value of feature data from multiple dimensions through weighted multi-dimensional risk algorithm , in order to assess the risk level of the regulated entity.

7. The big data-based intelligent supervision method according to claim 6 is characterized in that: The risk classification and strategy generation in S33 include: S331, Risk classification: Based on the calculation results of the multi-dimensional risk assessment algorithm , the comprehensive risk value is compared with the preset lower limit threshold and upper threshold Make comparisons and classify risks into different levels, including low risk, medium risk, and high risk; S332, strategy generation: According to different risk levels, corresponding regulatory strategies are automatically generated. For low risk levels, only information is recorded. For medium risk levels, warnings are issued and regulatory strategies are initiated. For high risk levels, emergency response procedures are triggered to adjust regulatory intensity. Regulatory strategies include adjustment of monitoring frequency, reallocation of resources, and triggering of risk response actions.

8. The intelligent supervision method based on big data according to claim 7 is characterized in that: The dynamic warning in S4 includes: S41, Dynamic warning triggering: Based on the graded risk assessment results and the generated supervision strategy, combined with the preset warning rules, it is automatically determined whether a dynamic warning needs to be triggered. The warning rules include response thresholds and strategies for different risk levels, including: Low risk warning rules: Trigger condition: When the risk value Less than the lower threshold When triggered; Early warning mechanism: Generate low-priority early warning notifications, including background logging, low-frequency email reminders, or marking the monitoring device as a state to be observed; Medium risk warning rules: Trigger condition: When the risk value Between lower threshold and upper threshold If a risk is detected, a medium-risk warning is triggered; Early warning mechanism: When a medium-risk early warning is triggered, a medium-priority early warning notification will be generated, and a reminder will be sent to the relevant responsible person via email or instant messaging tools, and the risk will be marked as an alert status on the monitoring platform; High-risk warning rules: Trigger condition: When the risk value Above upper threshold Triggered when a serious risk is detected; Early warning mechanism: When a high-risk early warning is triggered, a high-priority early warning will be generated, and an emergency alert will be issued to relevant personnel through multiple channels. At the same time, a red alert will be issued in the monitoring platform, requiring immediate intervention; S42, warning rules combined with strategies: Warning rules not only consider the risk level, but also combine the generated supervision strategy to determine the warning form and intensity. According to the generated supervision strategy, the warning rules dynamically adjust the alarm frequency, notification channels and response levels, including: Low risk warning: Low risk warning is only monitored regularly. When the detected risk assessment value is lower than the lower limit threshold When the system is in operation, it only indicates the existence of potential risks; Medium risk warning: Regularly monitor the risk source for medium risk warning, and increase the density of monitoring as the risk persists or increases; High-risk warning: Real-time monitoring of high-risk warnings and continuous assessment of the status of risk sources until the risk is resolved; S43, risk response advice provision: If a dynamic warning is triggered, corresponding risk response advice will be automatically generated and provided based on the risk level and the generated regulatory strategy, including strengthening monitoring, adjusting resource scheduling, and implementing emergency measures.

9. The intelligent supervision method based on big data according to claim 1 is characterized in that: The supervision strategy optimization in S5 includes: S51, Strategy optimization trigger: Based on the risk assessment results and dynamic early warning feedback information, the effectiveness of the current regulatory strategy is analyzed in real time. When it is found that the current regulatory strategy cannot cope with the risk, the regulatory strategy optimization process will be automatically triggered; S52, application of self-learning mechanism: through the self-learning mechanism, the Q-learning algorithm is used to continuously optimize and update the supervision strategy.

10. A smart supervision system based on big data, used to implement a smart supervision method based on big data as described in any one of claims 1 to 9, characterized in that: Includes the following modules: Multi-source data collection module: Through sensors and data collection devices deployed at multiple monitoring points, multi-source data, including environmental data, business process data and equipment status data, can be collected in real time; Data preprocessing and spatiotemporal feature extraction module: preprocess the collected multi-source data, including data denoising, missing value filling and outlier identification, and extract spatiotemporal features from the preprocessed multi-source data; Multidimensional risk assessment module: Based on the extracted spatiotemporal features, a multidimensional risk assessment algorithm is applied to assess the behavior and status of the regulated object, risks are graded according to the assessment results, and corresponding regulatory strategies are generated; Dynamic warning module: Automatically trigger dynamic warnings based on the graded risk assessment results and generated supervision strategies, combined with preset warning rules; Regulatory strategy optimization module: Based on risk assessment results and dynamic early warning feedback, the regulatory strategy is optimized and updated through a self-learning mechanism.

Images