Intelligent political data middle table and data processing method applied by intelligent political data middle table

Abstract

The invention discloses an intelligent government data center and a data processing method applied by the intelligent government data center, and relates to the technical field of digital government services, and the intelligent government data center comprises an item model module which is used for establishing a four-in-one item model of government service items, certificates, data and systems and realizing dynamic modeling and optimization of item association relationships; the digital file handling module is configured with a uniform interface API and a standardized engineering standard and supports cross-level, cross-region, cross-system and cross-department digital file handling circulation of government affair service items; the capability center module is used for providing service resources and capability supply for all service applications of the intelligent political data center; the operation monitoring module is used for collecting full-process data of the digital files and dynamically monitoring service efficiency indexes in combination with an intelligent monitoring algorithm; the collaborative service engine realizes cross-domain service collaborative handling and service resource intelligent scheduling; the invention provides a digital support platform for realizing intelligent reconstruction of a service process, and the standardization and collaboration level of government affair service is improved.

Description

Technical Field

[0001] The present application relates to the field of digital government service technology, and in particular to an intelligent government-digital center and a data processing method for its application. Background Art

[0002] With the deepening of digital government construction, government services are gradually developing in the direction of "one-stop service" and "cross-domain service". However, the existing technologies still have the following outstanding problems: 1. System silos and double-entry issues: There are generally incompatible interfaces between the business approval systems of vertical management departments (such as social security, medical insurance, civil affairs, etc.) and local government service platforms. After receiving the documents in the vertical management system, window staff need to manually re-enter the case data into the local unified application and acceptance system, resulting in reduced work efficiency and a "data gap" between the vertical management system and the local platform.

[0003] 2. Inefficient cross-domain services: The existing "cross-provincial services" and "city-wide services" mostly rely on point-to-point system docking or manual collection and agency models. The application materials need to be submitted repeatedly and the business process is lengthy. For example, cross-provincial enterprise relocation registration requires submitting materials back and forth between two places, which takes a long time to process and increases the cost of services for the public.

[0004] To sum up, existing technologies are unable to effectively realize the association and coordinated scheduling of different service elements (such as matters, certificates, data, and systems) at the government service level, resulting in inefficient cross-level, cross-regional, and cross-system business processing and insufficient data utilization, making it difficult to meet the core demand of "efficiently getting things done." Therefore, there is an urgent need for a digital support platform that can break through government data barriers and realize intelligent reconstruction of service processes, so as to improve the standardization and coordination of government services. Summary of the invention

[0005] In order to provide a digital support platform that can break down government data barriers and realize intelligent reconstruction of service processes, so as to improve the standardization and coordination of government services, this application provides a data processing method for an intelligent government-digital center and its applications.

[0006] In the first aspect, the invention objective of the present application is achieved by adopting the following technical solutions: An intelligent government and digital middle platform, comprising: The event model module is used to establish a four-in-one event model of government service events, certificates, data, and systems, and to achieve dynamic modeling and optimization of event relationships based on graph database technology; The digital document processing module is equipped with a unified interface API and standardized engineering standards to support the cross-level, cross-regional, cross-system, and cross-departmental digital document processing of government service items; The capability center module integrates digital public support capabilities and provides service resources and capability supply for all service applications of the intelligent government digital center platform; the service applications support unified service registration, scheduling and authorization management; Run the monitoring module to collect data from the entire digital document processing process, and dynamically monitor service performance indicators in combination with intelligent monitoring algorithms, triggering an early warning when the service performance indicators meet abnormal early warning conditions; The collaborative service engine, based on reinforcement learning algorithms, dynamically matches the optimal service path for each government service item, realizing cross-domain collaborative service processing and intelligent scheduling of service resources.

[0007] By adopting the above technical solution, a collaborative work platform is provided which is based on a collaborative service engine and is provided with a matter model module, a digital processing module, a capability center module and an operation monitoring module; specifically, through the "four-in-one" matter model of matters, certificates, data and systems, machine-readable and intelligent association of government service elements is realized, supporting the automatic flow of cross-level, cross-regional and cross-system business, and solving the problems of high manual coordination costs and low material reuse rate under the traditional model; through the unified digital processing standards and intelligent scheduling engine (i.e. collaborative service engine) of the digital processing module, "three-level collaboration" (remote acceptance-local approval-unified issuance of certificates) across provinces and cities can be realized, which is conducive to streamlining the materials, processing procedures and links in the process of government services and shortening business processing time; data notarization is realized based on the capability center module and the operation monitoring module. In actual application, the whole case processing can be realized through blockchain technology Process traceability, the scope of process traceability covers application, acceptance, approval and other business processing links, improves the completeness of regulatory data and the synchronization rate of cross-platform data; then based on the intelligent analysis of the operation monitoring module, it can monitor the multiple core performance indicators in the government service process (such as material return rate and one-time completion rate) in real time and efficiently monitor, improve the accuracy of abnormal event identification and abnormal warning response time; further, in actual application, through the standardized interface generation capability of the matter model module, the cost of vertical management system docking is greatly reduced (for example, the original 10 working days / matter is now 0.5 days), and the coverage rate of city-wide affairs is improved, supporting the implementation of the "four less and one faster" (less filling, less reporting, less running, less waiting, and faster processing) service. Therefore, this application provides a digital support platform that can break through government data barriers and realize intelligent reconstruction of service processes, greatly improving the standardization and coordination level of government services.

[0008] In a preferred example of the present application: the event model module includes: A model configuration unit is configured with a visual intelligent terminal, and obtains association rule configuration instructions based on the visual intelligent terminal to configure association rules of items, certificates, data, and systems, and establish a four-in-one item model of government service items, certificates, data, and systems; The graph analysis unit obtains the digital archive portrait of the government affairs generated by the service application and the digital archive of the digital handling module throughout its life cycle, and generates a departmental authority and responsibility graph and an item relationship graph in combination with the constructed four-in-one item model; the departmental authority and responsibility graph includes the authority and responsibility relationship context between different departments, and the item relationship graph includes the relationship between the items before and after the items; The iterative optimization unit simulates the impact of changes in the handling process of government service items based on AI algorithms, generates optimization plans based on the simulated impact results, and pushes them to the corresponding business department terminals.

[0009] By adopting the above-mentioned technical solutions, multi-source data fusion modeling and AI-driven process simulation can be realized. Specifically, a four-dimensional knowledge graph of "matters-certificates-data-system" is constructed through government digital archives (structured data) + digital processing logs (time series data) + visual configuration instructions (semi-structured data), which is conducive to improving the knowledge coverage and system configuration efficiency of the intelligent government digital platform; a reinforcement learning algorithm (DQN) is used to simulate the impact of process changes, achieve the accuracy of change risk prediction and generate optimization solutions; and by outputting executable process transformation suggestions, it is conducive to reducing labor costs, such as merging three high-frequency serial matters into one comprehensive service.

[0010] In a preferred example of the present application: the unified interface API includes a unified document handling interface standard; the digital document handling module includes: The standard generation unit generates standardized engineering standards and cross-system unified handling interface standards based on the four-in-one matter model; The circulation control unit uses blockchain to store and circulate the application data of the service, so as to achieve the traceability of the entire process and prevent the tampering of the data; The adapter unit connects to the vertical management system interface based on the unified cross-system document handling interface standard to complete cross-system data format conversion and field mapping.

[0011] By adopting the above technical solutions, the four-in-one matter model automatically generates interface specifications that comply with the GB / T 39786-2021 standard, thereby improving interface compatibility and optimizing development efficiency. For example, the time required to generate interface documents is reduced from 3 days to 2 hours; blockchain evidence technology builds a full-link evidence system of "application materials-approval records-completion results" to realize data tampering detection; the adapter unit supports protocol conversion of 8 vertical management systems (such as taxation, social security, and provident fund), and the data mapping accuracy is relatively high.

[0012] In a preferred example of the present application: the capability hub module includes: The service aggregation unit configures and integrates digital public support service capabilities and third-party service capabilities at different regional levels to support data collection and data flow of multimodal government service data; The service management unit obtains and displays the service resources and capability lists of all service applications of the intelligent government data center, and obtains and updates service resources based on the received service management configuration instructions; An intelligent scheduling engine that uses a weighted round-robin algorithm to allocate optimal service resources for current government services based on the service resource parameters of the service application; the service resource parameters include service response time, concurrent capacity, and security level; The permission management unit supports fine-grained permission control and realizes dynamic authentication of cross-region service calls.

[0013] By adopting the above technical solutions, dynamic scheduling based on weighted polling algorithm (for example, response time weight 50%, concurrent capacity 30%, security level 20%) is conducive to improving resource utilization and optimizing service availability; the four-dimensional authority matrix (role / area / matter / data level) is conducive to improving authority coverage and security audit efficiency, and the intelligent scheduling engine enables the service resource pool of the intelligent government data center to support dynamic expansion in a short period of time to cope with sudden traffic, such as social security supplementary payment peaks, which is conducive to improving the service governance efficiency of government service matters.

[0014] In a preferred example of the present application: the operation monitoring module includes: The performance analysis unit determines a number of service performance indicators for evaluating the completion of an event, and performs service performance analysis based on the service performance indicators to obtain service performance analysis results, wherein the service performance analysis results include the time consumption for event handling, the material reuse rate, and the one-time completion rate; The anomaly detection unit, based on time series analysis and machine learning models, identifies and prompts abnormal events in the current processing cycle of each government service item; A prediction and early warning unit is used to establish a historical data training prediction model, in which the risk of service interruption at the middle office and the risk of service abnormality of each government service item within a specified time period are predicted; and an early warning prompt is triggered when the risk of service interruption or abnormality at the middle office meets the preset early warning judgment trigger condition; The visualization display module provides a real-time monitoring view of the operating status of the intelligent government and digital center; the real-time monitoring view includes but is not limited to service performance analysis results.

[0015] By adopting the above technical solutions, multimodal anomaly detection is conducive to improving the accuracy of anomaly identification, the prediction and warning unit improves the abnormal emergency response capability, and real-time performance monitoring is achieved through the visualization module. The time for discovering deviations in policy implementation is shortened from months to weeks, which indirectly improves public satisfaction.

[0016] In the second aspect, the invention objective of the present application is achieved by adopting the following technical solutions: A data processing method applied to an intelligent government data center, comprising: Use graph database technology to build a four-in-one model of government service items, certificates, data, and systems, dynamically model and optimize the relationship between items, and obtain a standardized relationship network for corresponding government service items; Configure unified interface APIs and formulate standardized engineering standards based on government service items to achieve digital document circulation across levels, regions, systems and departments; A capability center module is configured to integrate cross-level public support capabilities and third-party service capabilities. The capability center module provides unified service registration, scheduling and authorization management for all service applications. Collect data from the entire digital document processing process, and dynamically monitor service performance indicators in combination with intelligent monitoring algorithms. When the service performance indicators reach the preset abnormal warning conditions, trigger an early warning. A reinforcement learning algorithm is used to dynamically match the optimal service path according to the characteristics of each government service item, and a collaborative service engine is used to achieve cross-domain service collaboration and intelligent scheduling of service resources.

[0017] By adopting the above technical solutions, the graph database constructs a heterogeneous knowledge network of government service items, certificates, data, and systems, and realizes structured mapping of multi-source data, which is conducive to improving data fusion efficiency and data integration capabilities; the dynamic topology algorithm supports real-time calculation of item weight coefficients, and the unified API interface specification reduces the cost of cross-departmental system docking; the unified scheduling strategy reduces the idle rate of service resources, that is, optimizes the utilization rate of service resources, enhances the platform's security management and control capabilities, and strengthens the learning algorithm to dynamically match the optimal service path. The average number of user processing steps is reduced, which is conducive to improving the one-time completion rate of items.

[0018] In a preferred example of the present application: the graph database adopts a heterogeneous graph storage architecture to construct a matter knowledge graph, a license relationship graph, a data resource graph, and a system interaction graph respectively; The dynamic topology optimization algorithm is used to calculate the weight coefficient of the event association relationship, and the weight coefficient of the event association relationship is updated based on the preset weight update frequency to achieve adaptive adjustment of the event association relationship; In the service application, key government service items and corresponding key item nodes are determined, and a semantically enhanced index is established in the key event nodes to support a hybrid retrieval mode of natural language query and graph retrieval.

[0019] By adopting the above technical solutions, the hybrid retrieval performance of the heterogeneous graph storage architecture integrates semantic indexing with graph retrieval, thereby improving the accuracy of natural language queries; by adopting a heterogeneous graph storage architecture to build a variety of graphs (matter knowledge graph, license and certificate relationship graph, data resource graph, and system interaction graph), it is possible to achieve refined modeling of complex relationships in government services and improve information retrieval efficiency and accuracy; in actual applications, for hot issues in government service matters, high-frequency matters (such as social security processing) can be automatically identified through weight fluctuation analysis, thereby improving the accuracy of priority scheduling, which is conducive to improving the digital processing efficiency of the intelligent government and digital platform.

[0020] In a preferred example of the present application, the service performance indicator is dynamically monitored, and an early warning is triggered when the service performance indicator reaches a preset abnormal early warning condition, including: Collect structured data, semi-structured data and unstructured data in the entire process of digital document processing; Use a multimodal data fusion engine to convert different data sources into a unified time-series data stream and build a three-dimensional feature matrix based on service requests, system responses, and resource scheduling; Based on the item association relationship in the graph database, a data anomaly propagation model is established. When the service performance index reaches the preset abnormal warning condition, an alarm is triggered. A causal reasoning engine is used in combination with the three-dimensional feature matrix to perform root cause analysis, determine all service items and their upstream and downstream systems affected by the corresponding government service items, and obtain abnormal impact analysis results; The warning level is determined based on the abnormal impact analysis result, the automatic processing strategy is triggered based on the warning level, and the process-related data is recorded through the blockchain evidence structure to obtain abnormal event processing traceability information.

[0021] By adopting the above technical solutions, multimodal data fusion is conducive to subsequent data analysis. Combining the data anomaly propagation model with the causal reasoning engine for root cause analysis can not only timely discover abnormalities in service performance indicators, but also deeply explore the specific causes of the anomalies and their scope of influence, providing a basis for accurate processing; the blockchain evidence storage structure records process-related data, which is conducive to achieving transparency and traceability of the abnormal event handling process.

[0022] In a preferred example of the present application, the service application is associated with a regional location identifier, the government service item is associated with a service item identifier, and the method further includes: Adopting an adaptive threshold learning algorithm, the warning threshold is dynamically calculated based on historical service performance data and real-time traffic characteristics of service applications; the warning threshold is set with multiple different warning levels, and different warning levels correspond to different handling strategies; Based on the regional location identifier and the service item identifier, combined with the spatiotemporal correlation analysis algorithm, when the service performance index reaches the preset abnormal warning condition, the regional location and government service item corresponding to the abnormal service performance are determined, and the warning level of the current abnormal event is adjusted.

[0023] By adopting the above technical solutions, the adaptive threshold learning algorithm can dynamically adjust the warning threshold based on historical data and real-time traffic characteristics of service applications, thereby improving the adaptability and sensitivity of the warning mechanism. Combined with the regional location identification and service item identification, when service performance abnormalities are detected, the specific affected areas and service items can be quickly located, which is conducive to taking targeted measures later.

[0024] In a preferred example of the present application, the causal reasoning engine is combined with the three-dimensional feature matrix to perform root cause analysis, determine all service items and their upstream and downstream systems affected by the corresponding government service items, and obtain abnormal impact analysis results, including: Using the relationship between items and service paths in the graph database, combined with the service request, system response and resource scheduling information extracted from the three-dimensional feature matrix, we can determine the current abnormal government service items and upstream and downstream dependencies. Utilize a causal reasoning engine to analyze the three-dimensional feature matrix, the current abnormal government service items, and the upstream and downstream dependencies, locate the specific propagation link that causes the abnormality, and determine all related service items and their upstream and downstream systems of the current abnormal government service items; Obtaining abnormal impact analysis results based on the reasoning results of the causal reasoning engine; and / or, Combined with the service path matching results in the reinforcement learning algorithm, multiple optional disposal plans corresponding to the abnormal impact analysis results are generated; The preset digital twin simulation module is used to simulate and deduce abnormal scenarios in a virtual environment, evaluate the treatment effects of multiple optional treatment plans, and determine the optimal treatment plan.

[0025] By adopting the above-mentioned technical solution, this application determines the specific propagation link of the anomaly through causal reasoning, and generates multiple optional disposal plans in combination with reinforcement learning. At the same time, it uses digital twin simulation to evaluate the effectiveness of the plan and finally selects the optimal disposal plan. This not only improves the scientificity and effectiveness of problem solving, but also can pre-verify the feasibility of the solution in a virtual environment and reduce actual operational risks.

[0026] In summary, the present application includes at least one of the following beneficial technical effects: 1. Provide a collaborative work platform based on a collaborative service engine, which is equipped with a matter model module, a digital case handling module, a capability center module, and an operation monitoring module; specifically, through the "four-in-one" matter model of matters, certificates, data, and systems, the machine-readable and intelligent association of government service elements is realized, supporting the automatic flow of cross-level, cross-regional, and cross-system business; 2. This application adopts a heterogeneous graph storage architecture to construct a variety of graphs (such as matter knowledge graph, certificate relationship graph, data resource graph and system interaction graph) to achieve refined modeling of complex relationships in government services and improve the efficiency of information retrieval. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is an architecture diagram of an intelligent political and digital middle platform in one embodiment of the present application; Figure 2 This is a diagram showing an example of the actual application of a matter model in an intelligent political and numerical center in one embodiment of the present application; Figure 3 This is a diagram showing an actual application example of a digital document processing module in an intelligent government and digital center in one embodiment of the present application; Figure 4 This is a diagram showing an example of the actual application of a capability center module of an intelligent political and digital center in one embodiment of the present application; Figure 5 It is a flow chart of a data processing method applied to an intelligent government data center in one embodiment of the present application. DETAILED DESCRIPTION

[0028] The present application is further described in detail below in conjunction with the accompanying drawings.

[0029] In one embodiment, if Figure 1 As shown, the present application discloses an intelligent government and digital middle platform, which includes a matter model module, a digital handling module, a capability center module, an operation monitoring module and a collaborative service engine. The detailed description of each functional module is as follows: like Figures 2 to 4As shown in the figure, the item model module is used to establish a four-in-one item model of government service items, certificates, data, and systems, and realize dynamic modeling and optimization of item association relationships based on graph database technology; specifically, the item model module includes a model configuration unit, a graph analysis unit, and an iterative optimization unit; the model configuration unit is configured with a visual intelligent terminal, and association rule configuration instructions are obtained based on the visual intelligent terminal to configure association rules for items, certificates, data, and systems, and establish a four-in-one item model of government service items, certificates, data, and systems; the visual intelligent terminal deployment supports a visual configuration interface for multi-role collaboration, including an item attribute editing area, a certificate association matrix diagram, and a data interface configuration card; the intelligent government and digital middle platform integrates a natural language processing module, and supports the generation of structured configuration parameters through text instructions, such as "forcibly associating enterprise establishment items with business licenses, legal person information databases, and market supervision systems", and setting a dynamic weight regulator to allow configuration personnel to adjust the association strength according to business priorities, such as increasing the weight of high-frequency items to 1.5 times; for example, React+Ant Design is used to build the front-end interface, and Spring is used to build the back-end interface. The Cloud microservice architecture enables real-time interaction between the rule engine and the graph database (Neo4j).

[0030] The graph analysis unit obtains the digital archive portrait of government affairs generated by the service application and the digital archive of the entire life cycle of digital affairs generated by the digital handling module, and generates a departmental authority and responsibility map and an issue relationship map in combination with the constructed four-in-one issue model; the departmental authority and responsibility map contains the authority and responsibility relationship context between different departments, and the issue relationship map contains the pre- and post-issue relationships of issues; the iterative optimization unit simulates the impact of changes in the handling process of government affairs service items based on the AI ​​algorithm, generates an optimization plan based on the simulated impact results and pushes it to the corresponding business department terminal. When simulating the government affairs service process, the process nodes include elements such as items, departments, and system interfaces; a reinforcement learning algorithm (such as DQN) is introduced, with the average handling time and the number of cross-department calls as optimization goals, and simulation optimization is performed based on a pre-set risk threshold. The simulation is automatically terminated when the risk threshold is exceeded, such as when the abnormality rate caused by the process change is >5%; when generating an optimization plan, strong correlation optimization and weak correlation glass can be performed, such as merging high-frequency related items (such as: "social security payment + electronic invoice printing" integrated into a comprehensive window) and relieving redundant system calls.

[0031] like Figure 3As shown, the digital document processing module is equipped with a unified interface API and standardized engineering standards to support the cross-level, cross-regional, cross-system, and cross-departmental digital document processing of government service items; standardized engineering standards refer to a set of specifications formulated to ensure the consistency of data interaction between different systems; the unified document processing interface standard defines the interface format and rules that need to be followed when exchanging data across systems; the unified interface API includes a unified document processing interface standard. Specifically, the digital document processing module includes a standard generation unit, a flow control unit, and an adaptation unit. The standard generation unit generates standardized engineering standards and cross-system unified document processing interface standards based on the four-in-one item model; the flow control unit uses blockchain evidence technology to store and circulate the service application's document processing data, realizing the traceability of the entire document processing process and the tamper-proofing of the document processing data. The tamper-proofing of case data adopts a cross-system and cross-departmental transfer permission control method, that is, a combination of dynamic permission management that defines a role permission matrix and a time window control method. For example, in service applications at the provincial and municipal levels, it is set that district-level approvers can access the "preliminary review" field, and municipal-level reviewers can access the "final review" field. The time window control is to set the field access validity period through a smart contract. For example, a "scanned copy of the business license" can only be called within 30 days after the application; a blockchain browser plug-in is developed to support data traceability query functions, such as tracing the entire process data by event ID, such as associating 5 system call records with the "business establishment" event, and statistically analyzing case efficiency indicators by time intervals. For example, the average processing time in a certain district in 30 days has been reduced from 8.2 days to 6.5 days; and trusted timestamp certification documents that meet the requirements of the "Electronic Signature Law" are exported.

[0032] The adapter unit connects to the vertical management system interface based on the cross-system unified document handling interface standard, and completes cross-system data format conversion and field mapping; the vertical management system interface refers to the API interface provided to the outside by the business system used by a specific vertical management department; data format conversion and field mapping refer to adjusting the data format from different systems to make it conform to the input requirements of the target system, and establishing a correspondence between the source field and the target field; specifically, the vertical management system connection and data conversion include interface adapter design and set data verification and fault-tolerant methods; for example, using JSON Schema to verify request or response data for structured verification; for unstructured data, OCR recognition + CV algorithm is used to detect the compliance of uploaded materials, such as whether the business license is within the validity period; when 5 consecutive interface calls time out (such as >3 seconds), the backup node is automatically switched and an alarm is triggered.

[0033] The capability center module is used to integrate digital public support capabilities and provide service resources and capability supply for all service applications of the intelligent government and digital middle platform; service applications support unified service registration, scheduling and authorization management; the capability center module includes a service aggregation unit, a service management unit, an intelligent scheduling engine and a permission management unit, among which the service aggregation unit is used to configure and integrate digital public support service capabilities and third-party service capabilities at different regional levels, and support data collection and data flow of multimodal government service data. Multimodal government service data includes structured, semi-structured and unstructured data forms; the service management unit is used to obtain and display the service resource and capability list of all service applications of the intelligent government and digital middle platform, and obtain and update service resources based on the received service management configuration instructions; the intelligent scheduling engine uses a weighted polling algorithm to allocate the optimal service resources for the current government service items based on the service resource parameters of the service application. The weighted polling algorithm is a scheduling algorithm that dynamically adjusts the request allocation weights based on multiple service quality indicators; service resource parameters include service response time, concurrent capacity and security level.

[0034] In this embodiment, the dynamic weight calculation formula of the weighted polling algorithm is: Among them, i is the service quality indicator, is the weight of the service quality indicator, t is the real-time service time; is the real-time response time (unit: ms); The current number of concurrent requests; is the safety level factor; , , is the weight coefficient, recommended value: =0.5, =0.3, =0.2 (SS level = 3, A level = 2, B level = 1); the embodiment of the present application sets the safety level coefficient as SS level = 3, A level = 2, B level = 1; the real-time scheduling decision is: when Less than 1s or Less than 50, it will be allocated to SS-level resources first. Greater than 1s and less than 5s or If it is greater than 50 and less than 200, it will be allocated to Class A resources according to the weight; If the time is greater than 5s, the circuit breaker mechanism is triggered and the backup resource pool is automatically switched.

[0035] The permission management unit configures service applications to support fine-grained permission control and realizes dynamic authentication of cross-regional service calls; for example, a dynamic authentication model based on fine-grained permission control is set up, and the dynamic authentication model includes a four-dimensional permission matrix. The four-dimensional permission parameters include roles, regions, matters and data levels, where roles include creators, approvers, reviewers, and the public based on RBAC role divisions; regions include provincial, municipal, and district and county levels to support cross-regional permission management; matters include business start-up, social security payment, and real estate registration based on business areas; data levels include full, summary, desensitized, and sensitive based on the scope of data visibility.

[0036] The operation monitoring module is used to collect the full-process data of digital processing, and dynamically monitor the service performance indicators in combination with the intelligent monitoring algorithm, and trigger an alert when the service performance indicators meet the abnormal warning conditions; the operation monitoring module includes an efficiency analysis unit, an abnormal detection unit, a prediction and warning unit, and a visualization display module; specifically, the efficiency analysis unit is used to determine a number of service performance indicators for the completion assessment of the matter, and perform service performance analysis based on a number of service performance indicators to obtain service performance analysis results, which include the time consumption of matter processing, material reuse rate, and one-time completion rate; the time consumption of matter processing is obtained by obtaining the time consumption of each node in the entire process of each matter (such as: acceptance → preliminary review → review → completion) through the API gateway embedding point, and calculating the average time consumption and standard deviation; the material reuse rate is to compare the application materials of different matters, and use the TF-IDF algorithm to calculate the material similarity, and the reuse rate = (number of similar materials / total number of materials) × 100%; the one-time completion rate refers to the statistical proportion of cases that do not require supplementary materials, and the counter is updated in real time through the Kafka message queue.

[0037] The anomaly detection unit identifies and prompts abnormal events in the current processing cycle of each government service item based on time series analysis and machine learning models; the anomaly detection unit adopts a hybrid detection model, which includes the isolation forest algorithm, the LSTM-Autoencoder algorithm and the random forest algorithm. The isolation forest algorithm is suitable for outlier detection, such as a sudden long processing time; the LSTM-Autoencoder algorithm is suitable for time series anomalies, such as a sudden increase in interface calls; the random forest algorithm is suitable for classification anomaly scenarios, such as abnormal material missing rate.

[0038] The prediction and early warning unit has established a historical data training prediction model. In the historical data training prediction model, the risk of service interruption in the middle office and the risk of service abnormality of each government service item within a specified time period are predicted; and the early warning prompt is triggered when the risk of service interruption or abnormality in the middle office meets the preset early warning judgment trigger conditions; specifically, the historical data training prediction model includes a time series prediction model and a gradient boosting tree model. The time series prediction model, such as the Prophet model, is used to predict the probability of service interruption in the next 30 days; the gradient boosting tree model predicts the abnormal risk of items, such as the increase in the risk value of "enterprise cancellation" items at the end of the quarter; the visualization display module is used to provide a real-time monitoring view of the operating status of the intelligent government digital middle office; the real-time monitoring view includes but is not limited to the service performance analysis results; the service performance analysis results refer to the service performance report calculated based on multiple service performance indicators, reflecting the overall performance of the current government services.

[0039] The collaborative service engine, based on reinforcement learning algorithms, dynamically matches the optimal service path for each government service item, realizing cross-domain collaborative service processing and intelligent scheduling of service resources.

[0040] Each module in the above-mentioned intelligent political and digital center can be fully or partially implemented through software, hardware and their combination; each module can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the corresponding operations of each module above.

[0041] Technical personnel in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the intelligent political and digital middle platform is divided into different functional units or modules to complete all or part of the functions described above.

[0042] In one embodiment, if Figure 5 As shown, a data processing method applied to an intelligent political and digital middle station is provided. The data processing method applied to the intelligent political and digital middle station is applied to an intelligent political and digital middle station. The data processing method applied to the intelligent political and digital middle station specifically includes the following steps: S1: Use graph database technology to build a four-in-one model of government service items, certificates, data, and systems, dynamically model and optimize the relationship between items, and obtain a standardized relationship network of corresponding government service items.

[0043] In this embodiment, a four-in-one item model is constructed using a graph database tool, such as Neo4j, which supports dynamic updates of model relationships, such as adding new certificate types or data interfaces; a standardized relationship network refers to a clear network structure of the relationships between service items and their associated certificates, data, and systems formed through modeling.

[0044] Specifically, before building the four-in-one matter model, first use the RPA tool to extract the vertical management system data, identify high-frequency bottlenecks (such as secondary entry, repeated submission of materials) through interviews, and then define the node types in the four-in-one matter model (matters, certificates, data, systems), and establish model relationship edges: matters → certificates that need to be submitted, matters → dependent data, and matters → processing systems.

[0045] Furthermore, the graph database adopts a heterogeneous graph storage architecture to construct a matter knowledge graph, a license relationship graph, a data resource graph and a system interaction graph respectively; the matter knowledge graph includes matter attributes and process information, the license relationship graph includes license types and related matters, the data resource graph includes data fields and sources, and the system interaction graph includes system interfaces and dependencies. Different entity relationships are distinguished by node types (such as matters, licenses, data, systems) and edge types (such as associations, dependencies, and calls), and multimodal data fusion is supported. For example, in the "second-hand house transfer" matter, the "real estate certificate" (license relationship graph), "real estate transaction data" (data resource graph) and "real estate registration system" (system interaction graph) are associated; a dynamic topology optimization algorithm is used to calculate the weight coefficient of the matter association relationship, and the matter association relationship weight coefficient is updated based on the preset weight update frequency to achieve matter association. Adaptive adjustment of relationships, where the dynamic topology optimization algorithm is based on the graph neural network (GNN) or PageRank algorithm, calculates the weight coefficients of the relationship between items (such as dependency strength, call frequency), and updates the weights at a preset frequency (such as daily / weekly) to achieve dynamic optimization of the relationship network; determines key government service items and corresponding key item nodes in the service application, establishes semantic enhanced indexes in key event nodes, and supports a hybrid retrieval mode of natural language query and graph retrieval; in this embodiment, a semantic index (such as BERT vector embedding) is established at the key nodes of the four-in-one item model (such as high-frequency items, core certificates and licenses), supports a hybrid mode of natural language query (such as "What materials do I need to transfer my household registration") and graph structured retrieval (such as "Query all items involving ID cards"), and combines NLP technology to parse user intent and map it to graph nodes and relationship paths.

[0046] S2: Configure a unified interface API based on government service items and formulate standardized engineering standards to achieve digital document flow across levels, regions, systems and departments.

[0047] In this embodiment, the unified interface API and the formulation of standardized engineering standards refer to the formulation of digital document processing interface specifications (such as RESTful API) across levels / regions / systems / departments, and the unified document processing data format (such as JSON-LD); protocol conversion, authentication and flow control are achieved through the service gateway (API Gateway).

[0048] S3: Configure a capability hub module that integrates cross-level public support capabilities and third-party service capabilities. The capability hub module provides unified service registration, scheduling, and authorization management for all service applications.

[0049] In this embodiment, the capability hub module can integrate provincial and municipal public support capabilities (such as electronic certificate libraries, unified identity authentication) and third-party services (such as logistics and express delivery), provide unified access and management, and based on the microservice architecture, realize dynamic online and offline capabilities and authority control through the service registration center. For example, when a company is established, it automatically calls the electronic seal service to sign documents and sends the business license through the logistics interface.

[0050] S4: Collect data from the entire digital processing process, and dynamically monitor service performance indicators using intelligent monitoring algorithms. When service performance indicators reach preset abnormal warning conditions, an early warning will be triggered.

[0051] In this embodiment, a visual intelligent terminal with a visual monitoring screen is built based on Prometheus+Grafana. The visual intelligent terminal supports custom alarm rules, such as a warning of a timeout in handling a case.

[0052] Specifically, real-time data on the entire digital processing process (such as processing time and material missing rate) is collected and combined with a threshold algorithm to trigger abnormal warnings; if the average processing time for a matter exceeds a preset threshold (such as 5 working days), it will be automatically pushed to government affairs management personnel for processing.

[0053] S5: Using reinforcement learning algorithms, the optimal service path is dynamically matched according to the characteristics of each government service item, and cross-domain service collaboration and intelligent scheduling of service resources are achieved through a collaborative service engine.

[0054] In this embodiment, a reinforcement learning algorithm such as DQN is used, and the optimal service path such as the fastest channel and the least material is used, and the DQN model is trained using historical case processing data.

[0055] Specifically, the path optimization of the reinforcement learning algorithm can be achieved by constructing the state space (such as the current processing progress), action space (optional service nodes), and reward function (efficiency first or minimum material).

[0056] In one embodiment, in step S4, the service performance indicator is dynamically monitored, and when the service performance indicator reaches a preset abnormal warning condition, a warning is triggered, including: S41: Collect structured data, semi-structured data and unstructured data in the entire process of digital document processing.

[0057] In this embodiment, structured data refers to data with a fixed format, such as event configuration parameters; semi-structured data refers to data with a certain structure but not completely regular, such as API call logs; unstructured data refers to data without a fixed format, such as user operation videos.

[0058] S42: Use the multimodal data fusion engine to convert different data sources into a unified time-series data stream and build a three-dimensional feature matrix based on service requests, system responses, and resource scheduling.

[0059] In this embodiment, the three-dimensional feature matrix refers to a time series data structure that represents information on service requests, system responses, and resource scheduling. Apache Kafka is used for data stream processing, and TensorFlow is used for feature engineering. Service requests include average service response time and timeout rate. System responses include CPU utilization, error code distribution, and API call success rate. Resource scheduling includes memory occupancy and the number of active threads in the thread pool.

[0060] S43: Based on the relationship between items in the graph database, a data anomaly propagation model is established. When the service performance index reaches the preset abnormal warning condition, an alarm is triggered. The causal reasoning engine is combined with the three-dimensional feature matrix to perform root cause analysis, determine all service items and their upstream and downstream systems affected by the corresponding government service items, and obtain the abnormal impact analysis results.

[0061] In this embodiment, the data anomaly propagation model refers to a model that describes how data anomalies propagate within the system; the causal reasoning engine is a tool for identifying the causes of specific events; the preset abnormal warning conditions include a data comparison method for dynamic warning thresholds and an abnormal node detection method for abnormal events; that is, based on the relationship between items in the graph database (such as upstream and downstream system call chains), an abnormal propagation path model is established, and the root cause is located in combination with a causal reasoning engine (such as DoWhy); root cause analysis refers to locating the cause of the abnormality and the impact of the abnormality when an abnormality occurs, using the graph attention network to calculate the node influence and screen key influencing factors.

[0062] Specifically, step S43 includes: S431: Utilize the matter association relationship and service path relationship in the graph database, combined with the service request, system response and resource scheduling information extracted from the three-dimensional feature matrix, to determine the current abnormal government service matters and upstream and downstream dependencies.

[0063] Specifically, the relationship between matters includes cross-departmental collaborative matters, matter-level relationships, reference relationships between materials and matters, relationships between fields, and relationships between service forms; among them, cross-departmental collaborative matters include government services that are jointly handled by multiple departments, such as the handling of certificates and licenses for natural persons and legal persons in different places, and the integration of multiple evaluations for investment projects; matter-level relationships include matters that are divided into national, provincial, municipal, county, township, and village levels according to the service level, forming a hierarchical service network; the reference relationship between materials and matters refers to the reference relationship when specific materials need to be associated with matter handling, such as the "enterprise establishment" matter needs to reference business licenses, legal person ID cards and other materials, which can be set through the result reference relationship and name reference relationship between matters and materials; the relationship between fields refers to the formation of field associations through data element splitting; the relationship between service forms refers to the distinction between online handling, offline handling, and online and offline integrated matters, for example, "one-stop service" requires the integration of multi-platform service entrances; the service path relationship includes cross-level flow paths, process integration paths, information sharing paths, and digital platform paths.

[0064] S432: Use the causal reasoning engine to analyze the three-dimensional feature matrix, the current abnormal government service items and upstream and downstream dependencies, locate the specific propagation link that causes the abnormality, and determine all related service items and their upstream and downstream systems of the current abnormal government service items.

[0065] Specifically, the causal reasoning engine is based on Bayesian networks or deep learning methods and is trained using historical data. It can handle complex relationships in three-dimensional feature matrices and accurately identify the root causes and impacts of government service items. Abnormal propagation link analysis refers to the use of a causal reasoning engine to analyze current abnormal government service items and their upstream and downstream dependencies, determine specific propagation links, for example, "Item A → System B → Interface C", and clarify how the abnormality propagates from one link to another.

[0066] S433: Obtain abnormal impact analysis results based on the reasoning results of the causal reasoning engine.

[0067] Specifically, the inference results include the results obtained based on the causal network and the quantification of causal effects between variables. The causal network is used to clarify the upstream and downstream system call chains and the abnormal propagation paths; for example, the abnormality of a certain matter may be related to the database service or pre-approval system it depends on, and the quantification of causal effects between variables includes quantifying the causal effects between variables such as service requests, system responses, and resource scheduling. For example, a 4s delay in the platform's system response time may cause the user complaint rate to increase by 5%; the abnormal impact analysis results include a detailed list of all matters and service paths affected by the abnormality, providing the root cause of the problem and solution suggestions.

[0068] and / or, S434: Combine the service path matching results in the reinforcement learning algorithm to generate multiple optional disposal plans corresponding to the abnormal impact analysis results.

[0069] Specifically, service path matching can be performed based on the state-action mapping relationship, path selection priority and environmental adaptability evaluation. For example, service path matching based on the state-action mapping relationship refers to clarifying the set of feasible actions under different service scenarios, such as: discrete action: selecting a specific service node or switching service path; continuous action: dynamically adjusting resource allocation weights or path priorities, and environmental adaptability evaluation analyzes the robustness and generalization ability of path strategies in different scenarios; based on the service path matching results, multiple optional disposal plans are generated, such as automatically switching to backup nodes, prioritizing resource scheduling, etc.

[0070] S435: Use the preset digital twin simulation module to simulate and deduce abnormal scenarios in a virtual environment, evaluate the treatment effects of multiple optional treatment plans, and determine the optimal treatment plan.

[0071] Specifically, the digital twin simulation module is a tool that can simulate the actual system operation in a virtual environment; reproduce the current abnormal scenario in the digital twin environment, import multiple optional disposal plans, and perform operation simulation: observe the performance of each plan in the virtual environment, and record key indicators (such as recovery time, resource utilization, etc.); then evaluate the disposal effect of each plan based on the simulation results and select the optimal disposal plan.

[0072] S44: Determine the warning level based on the abnormal impact analysis results, trigger the automated processing strategy based on the warning level, and record the process-related data through the blockchain evidence structure to obtain abnormal event processing traceability information.

[0073] In this embodiment, multiple different warning levels are set according to different levels of abnormal severity, such as local service interruption, cross-departmental flow blockage, etc.; automated processing strategies include service fusing, resource expansion, and generating warning processing work orders and assigning responsible persons; abnormal event processing traceability information includes the cause of the abnormal event, processing flow, impact scope, and abnormal time process information of the disposal results.

[0074] Specifically, the warning level (such as P0-P3) is divided according to the scope of abnormal impact and business importance, and the process data is stored through blockchain coarse evidence technology. Combined with rule engines (such as Drools) and smart contracts, automated decision-making is achieved. The specific automated processing strategy can be customized based on expert knowledge and experience.

[0075] In one embodiment, a data processing method applied to an intelligent political and digital middle station also includes: S10: Adopting adaptive threshold learning algorithm, the warning threshold is dynamically calculated according to the historical service performance data and the real-time traffic characteristics of the service application; the warning threshold is set with multiple different warning levels, and different warning levels correspond to different handling strategies.

[0076] In this embodiment, historical service performance data includes case handling success rate, response time, and one-time completion rate; real-time traffic characteristics include the number of concurrent requests and system load; the adaptive threshold learning algorithm is a dynamic threshold adjustment model based on reinforcement learning, which uses sliding windows and online learning algorithms to update thresholds in real time. Historical data shows that the normal response time for a certain matter is 50-100ms. When the real-time traffic surges, the algorithm automatically adjusts the upper threshold limit to 150ms; the disposal strategy refers to the specific response measures formulated for different warning levels, and supports differentiated strategy configuration for multiple warning levels (such as P0-P3); For example, the dynamic threshold is calculated as: response time threshold = baseline value × (1 + fluctuation coefficient).

[0077] S20: Based on the regional location identification and service item identification, combined with the spatiotemporal correlation analysis algorithm, when the service performance index reaches the preset abnormal warning condition, the regional location and government service item corresponding to the abnormal service performance are determined, and the warning level of the current abnormal event is adjusted.

[0078] In this embodiment, the regional location identifier is a label used to identify the physical or logical location of the service application, which is convenient for distinguishing the service performance of different regions; the service item identifier is a unique identifier used to identify a specific government service item, which helps to locate the specific business process; the spatiotemporal association analysis algorithm is a data analysis method that combines time and space dimensions, and is used to identify and analyze the spatial distribution and time patterns of abnormal events. It combines the regional geographical location (such as provincial and municipal administrative divisions, service network coordinates) and service item identifiers (such as item codes, affiliated departments), and analyzes the geographical distribution and diffusion trends of abnormal events through a spatiotemporal clustering algorithm (such as DBSCAN).

[0079] Specifically, we conduct multi-dimensional abnormal warning factor analysis based on spatiotemporal data, generate abnormal events and associate them with service items; the adjustment of warning levels includes calculating risk scores based on abnormal duration, impact range, and user complaints, triggering adjustment strategies: P0 (province-wide fuse) → P3 (local alarm) It should be understood that the serial numbers of the steps in the above embodiments do not imply a sequence of execution. The execution sequence of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0080] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, it should be understood by those skilled in the art that the technical solutions described in the aforementioned embodiments may still be modified, or some of the features thereof may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. An intelligent political and digital middle platform, characterized in that: include: The event model module is used to establish a four-in-one event model of government service events, certificates, data, and systems, and to achieve dynamic modeling and optimization of event relationships based on graph database technology; The digital document processing module is equipped with a unified interface API and standardized engineering standards to support the cross-level, cross-regional, cross-system, and cross-departmental digital document processing of government service items; The capability center module integrates digital public support capabilities and provides service resources and capability supply for all service applications of the intelligent government digital center platform; the service applications support unified service registration, scheduling and authorization management; Run the monitoring module to collect data from the entire digital document processing process, and dynamically monitor service performance indicators in combination with intelligent monitoring algorithms, triggering an early warning when the service performance indicators meet abnormal early warning conditions; The collaborative service engine, based on reinforcement learning algorithms, dynamically matches the optimal service path for each government service item, realizing cross-domain collaborative service processing and intelligent scheduling of service resources.

2. According to claim 1, the intelligent political and digital middle platform is characterized in that: The event model module includes: A model configuration unit is configured with a visual intelligent terminal, and obtains association rule configuration instructions based on the visual intelligent terminal to configure association rules of items, certificates, data, and systems, and establish a four-in-one item model of government service items, certificates, data, and systems; The graph analysis unit obtains the digital archive portrait of the government affairs generated by the service application and the digital archive of the digital handling module throughout its life cycle, and generates a departmental authority and responsibility graph and an item relationship graph in combination with the constructed four-in-one item model; the departmental authority and responsibility graph includes the authority and responsibility relationship context between different departments, and the item relationship graph includes the relationship between the items before and after the items; The iterative optimization unit simulates the impact of changes in the handling process of government service items based on AI algorithms, generates optimization plans based on the simulated impact results, and pushes them to the corresponding business department terminals.

3. According to claim 1, the intelligent political and digital middle platform is characterized in that: The unified interface API includes a unified document handling interface standard; the digital document handling module includes: The standard generation unit generates standardized engineering standards and cross-system unified handling interface standards based on the four-in-one matter model; The circulation control unit uses blockchain to store and circulate the application data of the service, so as to achieve the traceability of the entire process and prevent the tampering of the data; The adapter unit connects to the vertical management system interface based on the unified cross-system document handling interface standard to complete cross-system data format conversion and field mapping.

4. According to claim 1, the intelligent political and digital middle platform is characterized in that: The capability hub module includes: The service aggregation unit configures and integrates digital public support service capabilities and third-party service capabilities at different regional levels to support data collection and data flow of multimodal government service data; The service management unit obtains and displays the service resources and capability lists of all service applications of the intelligent government data center, and obtains and updates service resources based on the received service management configuration instructions; An intelligent scheduling engine that uses a weighted round-robin algorithm to allocate optimal service resources for current government services based on the service resource parameters of the service application; the service resource parameters include service response time, concurrent capacity, and security level; The permission management unit supports fine-grained permission control and realizes dynamic authentication of cross-region service calls.

5. According to claim 1, the intelligent political and digital middle platform is characterized in that: The operation monitoring module includes: The performance analysis unit determines a number of service performance indicators for evaluating the completion of an event, and performs service performance analysis based on the service performance indicators to obtain service performance analysis results, wherein the service performance analysis results include the time consumption for event handling, the material reuse rate, and the one-time completion rate; The anomaly detection unit, based on time series analysis and machine learning models, identifies and prompts abnormal events in the current processing cycle of each government service item; A prediction and early warning unit is used to establish a historical data training prediction model, in which the risk of service interruption at the middle office and the risk of service abnormality of each government service item within a specified time period are predicted; and an early warning prompt is triggered when the risk of service interruption or abnormality at the middle office meets the preset early warning judgment trigger condition; The visualization display module provides a real-time monitoring view of the operating status of the intelligent government and digital center; the real-time monitoring view includes but is not limited to service performance analysis results.

6. A data processing method applied to an intelligent political and digital platform, characterized in that: include: Use graph database technology to build a four-in-one model of government service items, certificates, data, and systems, dynamically model and optimize the relationship between items, and obtain a standardized relationship network for corresponding government service items; Configure unified interface API and formulate standardized engineering standards based on government service items to realize digital document circulation across levels, regions, systems and departments; A capability center module is configured to integrate cross-level public support capabilities and third-party service capabilities. The capability center module provides unified service registration, scheduling and authorization management for all service applications. Collect data from the entire digital document processing process, and dynamically monitor service performance indicators using intelligent monitoring algorithms. When the service performance indicators reach preset abnormal warning conditions, trigger warnings. A reinforcement learning algorithm is used to dynamically match the optimal service path according to the characteristics of each government service item, and a collaborative service engine is used to achieve cross-domain service collaboration and intelligent scheduling of service resources.

7. According to claim 6, a data processing method applied to an intelligent political and digital middle station is characterized in that: The graph database adopts a heterogeneous graph storage architecture to construct a matter knowledge graph, a license relationship graph, a data resource graph, and a system interaction graph respectively; The dynamic topology optimization algorithm is used to calculate the weight coefficient of the event association relationship, and the weight coefficient of the event association relationship is updated based on the preset weight update frequency to achieve adaptive adjustment of the event association relationship; In the service application, key government service items and corresponding key item nodes are determined, and a semantically enhanced index is established in the key event nodes to support a hybrid retrieval mode of natural language query and graph retrieval.

8. A data processing method applied to an intelligent political and digital middle station according to claim 6 or 7, characterized in that: The dynamic monitoring of the service performance indicator and triggering of an early warning when the service performance indicator reaches a preset abnormal early warning condition include: Collect structured data, semi-structured data and unstructured data in the entire process of digital document processing; Use a multimodal data fusion engine to convert different data sources into a unified time-series data stream and build a three-dimensional feature matrix based on service requests, system responses, and resource scheduling; Based on the item association relationship in the graph database, a data anomaly propagation model is established. When the service performance index reaches the preset abnormal warning condition, an alarm is triggered. A causal reasoning engine is used in combination with the three-dimensional feature matrix to perform root cause analysis, determine all service items and their upstream and downstream systems affected by the corresponding government service items, and obtain abnormal impact analysis results; The warning level is determined based on the abnormal impact analysis result, the automatic processing strategy is triggered based on the warning level, and the process-related data is recorded through the blockchain evidence structure to obtain abnormal event processing traceability information.

9. A data processing method applied to an intelligent political and digital middle station according to claim 6 or 7, characterized in that: The service application is associated with a regional location identifier, and the government service item is associated with a service item identifier, and the method further includes: Adopting an adaptive threshold learning algorithm, the warning threshold is dynamically calculated based on historical service performance data and real-time traffic characteristics of service applications; the warning threshold is set with multiple different warning levels, and different warning levels correspond to different handling strategies; Based on the regional location identifier and the service item identifier, combined with the spatiotemporal correlation analysis algorithm, when the service performance index reaches the preset abnormal warning condition, the regional location and government service item corresponding to the abnormal service performance are determined, and the warning level of the current abnormal event is adjusted.

10. The data processing method applied to the intelligent political and digital middle station according to claim 8 is characterized in that: The causal reasoning engine is used in combination with the three-dimensional feature matrix to perform root cause analysis, determine all service items and their upstream and downstream systems affected by the corresponding government service items, and obtain abnormal impact analysis results, including: Using the relationship between items and service paths in the graph database, combined with the service request, system response and resource scheduling information extracted from the three-dimensional feature matrix, we can determine the current abnormal government service items and upstream and downstream dependencies. Utilize a causal reasoning engine to analyze the three-dimensional feature matrix, the current abnormal government service items, and the upstream and downstream dependencies, locate the specific propagation link that causes the abnormality, and determine all related service items and their upstream and downstream systems of the current abnormal government service items; Obtaining abnormal impact analysis results based on the reasoning results of the causal reasoning engine; and / or, Combined with the service path matching results in the reinforcement learning algorithm, multiple optional disposal plans corresponding to the abnormal impact analysis results are generated; The preset digital twin simulation module is used to simulate and deduce abnormal scenarios in a virtual environment, evaluate the treatment effects of multiple optional treatment plans, and determine the optimal treatment plan.

Images