A knowledge-driven spatio-temporal data collaborative linkage updating method

Abstract

The application discloses a kind of knowledge driving-based spatio-temporal data collaborative linkage update method, belong to data management technical field, method includes: the unified knowledge graph of fusing business rules, spatial relationship and time association is constructed, the quantitative characterization of entity multidimensional relationship is realized;Based on incremental change detection, significant change entity is identified, and the associated entity set is intelligently discovered using knowledge graph;Adopt event-driven model and priority scheduling to trigger linkage update;Through quantitative conflict detection and business priority-based resolution mechanism, data logic consistency is guaranteed;Realize cross-level synchronous monitoring and whole-process version management.The method overcomes the linkage difficulty and poor consistency of natural resources and other cross-domain multi-source spatio-temporal data due to model heterogeneity and implicit update rules, can effectively improve the automation, intelligent level and overall consistency of multi-source spatio-temporal data linkage update, suitable for land space planning, ecological protection and other scenarios requiring multi-department data collaboration.

Description

Technical Field

[0001] The present invention relates to the technical field of data management, and in particular to a knowledge-driven spatio-temporal data collaborative linkage updating method. Background Art

[0002] In many fields such as natural resource management, smart city and environmental monitoring, the core business relies on accurate management and dynamic updating of geospatial entities (such as plots, buildings, facilities) and their attribute states (such as ownership, use, status) evolving over time. These data usually come from different business departments such as planning, examination and approval, investigation and registration, forming multi-source heterogeneous data sets, and are often distributed in business systems at different levels such as national, provincial, municipal and county levels. Ideally, these data should form a "one-map" pattern with unified logic and dynamic updating, so as to provide an accurate and consistent base map for comprehensive decision-making.

[0003] However, in the practice of realizing dynamic linkage updating of "one-map" data, a series of technical challenges are faced. First, there are conflicts and implicit problems in data models and rules. Data models (such as plot coding systems and attribute structures) from different business departments have differences, and business rules that drive data updating are usually contained in regulations or manual experience, and have not been converted into explicit logic that can be directly understood and executed by computers. This results in that data updating is often limited to within a department, cross-departmental updating dependencies are unclear, and automatic triggering is difficult. Second, there are difficulties in change perception and influence scope identification. Data updating may originate from any business link, and the system needs to automatically and accurately detect which entities have undergone significant changes with business significance from massive continuously generated business logs and spatial data. More importantly, a change of one entity may affect other entities through various complex relationships such as spatial adjacency, business chains and time series. Traditional methods are usually processed through hard coding or simple preset association rules, which lack flexibility, are difficult to adapt to the adjustment of business rules, cannot accurately define the influence scope of complex network shape, and easily lead to update omission or excessive diffusion.

[0004] Data updates can lead to conflicts and consistency issues. When multiple update events occur concurrently or in a chain, they may propose mutually exclusive modifications to the same spatial region or logical entity, resulting in spatial or logical conflicts. Existing systems often rely on manual coordination, which is inefficient and inconsistent in standards. Furthermore, after completing a series of automated updates, there is a lack of effective automated evaluation methods to verify whether the updated dataset maintains the expected logical consistency and statistical characteristics. Cross-level synchronization and process traceability are also crucial issues. Update results need to be synchronized efficiently and reliably to nodes at all levels, but the completeness and timeliness of the synchronization process lack effective monitoring. In addition, the cause, process, and result of each linked update are not systematically recorded, making it difficult to quickly trace and locate problems when data anomalies occur or auditing is required.

[0005] Existing data integration and update technologies, such as batch synchronization based on ETL (Extract, Transform, Load), update notifications based on message middleware, or business collaboration based on workflows, have solved the problems of data transmission and simple linkage to some extent. However, they are usually unable to deeply address the core pain points of logical fragmentation and poor consistency caused by implicit business rules, complex relationships, and frequent conflicts. Therefore, there is an urgent need for a spatiotemporal data collaborative update method that can deeply integrate business knowledge, intelligently perceive changes, accurately infer impacts, automatically resolve conflicts, and ensure that the entire process is manageable and controllable. Summary of the Invention

[0006] This invention overcomes the problems of difficulty in linkage and poor consistency caused by heterogeneous models and implicit update rules in cross-domain multi-source spatiotemporal data such as natural resources. It provides a knowledge-driven spatiotemporal data collaborative linkage update method, which can significantly improve the automation, intelligence level and overall consistency of multi-source spatiotemporal data linkage update. It is applicable to scenarios that require multi-department data collaboration, such as land spatial planning and ecological protection.

[0007] To achieve the above objectives, the present invention adopts the following solution: A knowledge-driven spatiotemporal data collaborative update method includes the following steps: Step 1: Analyze the natural resource business process and extract business rules. Use OWL2 language to convert the business rules into machine-executable rules. Calculate the spatial adjacency of geographic entities and the temporal correlation of geographic entity states based on the standard defined topological relationship. With geographic entities as core nodes, machine-executable rules, adjacency, and temporal correlation as edge attributes, construct a business-space-temporal knowledge graph. Step 2: Based on log analysis, time series comparison and spatial overlay technology, extract the attribute change increment of structured business data and the spatial change increment of spatial data, and calculate the comprehensive change degree of geographic entities based on the attribute change increment and spatial change increment. When the comprehensive change degree exceeds the preset threshold, it is determined that the geographic entity has undergone significant change. Step 3: When a geographic entity undergoes significant changes, a breadth-first search is performed based on the business-spatial-temporal knowledge graph to identify the set of associated geographic entities related to the geographic entity, and an event-driven model is used to trigger the synchronous update of the set of associated geographic entities. Step 4: Calculate the spatial overlap between related geographic entities during the synchronous update process. When the spatial overlap exceeds the preset overlap threshold, it is determined to be a spatial conflict. The spatial conflict is resolved based on business priority, and the consistency of data distribution before and after the update is evaluated using relative entropy. Step 5: Calculate the time delay for each level of regional nodes to complete the synchronization update, and generate a version number containing metadata about the update initiation time, geographic entity status, and operator to support version rollback.

[0008] Preferably, in step 1, the business rules are converted into machine-executable rules using the OWL2 language, and the quantization function used for the rule triggering conditions is defined as follows: ; in, a k ( expected value ) is a business event expected value The k One attribute, V k This is the set of trigger thresholds corresponding to this attribute. m The total number of attributes associated with the rule; quantified. This is a Boolean value; when its value is 1, it indicates that the rule has been triggered.

[0009] As a preferred embodiment, in step 1, the specific method for calculating the spatial adjacency of geographic entities is as follows: Based on the standard-defined topological relationships, calculate the spatial entities in two geographic entities. S 1 and S The adjacency degree of 2 is calculated using the following formula: in, Representing spatial entities S The boundary, Indicates the boundary length; when adjacent degree When they are, they are determined to be adjacent geographical entities; The specific method for calculating the temporal correlation degree of geographic entity status is as follows: Calculate geographic entities eAt the point of time t 1 and t 2. State correlation The calculation formula is: in, λ This is the time decay coefficient, with a value ranging from 0.01 to 0.1; When constructing a business-spatial-temporal knowledge graph, geographic entities are used as core nodes, and machine-executable rules are incorporated. Adjacency and time correlation As edge attributes connecting geographic entity nodes, they are used in fusion modeling to obtain a unified representation of entities, relationships, and rules.

[0010] As a preferred method, in step 2, the specific way to extract the attribute change increment of structured business data is as follows: For structured business data, compare the hash values ​​of new and old fields to detect attribute changes. At that time, determine the attribute increment. ,in For hash functions, a new For new field data, a old This is old field data; The method for extracting spatial change increments from spatial data is as follows: Calculate two periods of spatial data S new and S old The symmetric difference between them yields the spatial variation increment. ΔS ,Right now: ,in Added data This is a reduction in the amount of data. Calculate the overall degree of change of geographic entities C ( e The specific formula is as follows: in, α This represents the spatial weight, with a value ranging from 0.6 to 0.8. Represents the area of ​​a space; indicator function In attribute increment The value is 1 if the condition is met, and 0 otherwise. w k For the first k Preset weights for each attribute, n Total number of attributes; The preset threshold for the overall degree of change is θ,when At that time, determine the geographic entity e Significant changes have occurred.

[0011] Preferably, in step 3, a breadth-first search is performed based on the business-space-temporal knowledge graph to identify the set of associated geographic entities related to the geographic entities. E rel The specific identification and judgment conditions are as follows: in, expected value To act on geographic entities e And business events that triggered significant changes in it, e' Indicates candidate associated geographic entities, conditions , , These correspond to spatial adjacency, temporal association, and business rule association between two geographical entities, respectively. Synchronous updates are triggered using an event-driven model, specifically including: For the set of associated geographic entities E rel Each geographic entity in e' Calculate update priority P ( e' The calculation formula is: in, β and γ The weighting coefficients and β+γ≤1 , The association weights are assigned based on business rules; Based on the update priority of each geographic entity, the synchronous update of the associated geographic entities is triggered and executed.

[0012] Preferably, in step 4, the geographic entity is calculated using the following formula. e 1 and e 2 Between: When spatial overlap At that time, it was determined to be a spatial conflict; The process of resolving spatial conflicts based on business priorities specifically includes: Calculate business priorities for conflicting geographic entities P bus ( e The calculation formula is: in, b k Representing geographic entitiese The associated first k Item business type, level ( b k This indicates the preset level for this service type. w k The preset weight for this business type, m This represents the total number of associated business types. When conflicts are resolved, the update results of geographic entities with higher business priority will override the update results of geographic entities with lower business priority.

[0013] Preferably, in step 4, relative entropy is used to evaluate the consistency of the data distribution before and after the update, specifically including: Calculate the data distribution before the update P Compared with the updated data distribution Q The relative entropy between : in, N The total number of values ​​that the data can take; when If the value is less than the preset consistency threshold, the update is deemed to meet the consistency requirements.

[0014] Preferably, in step 5, the time delay for each level of node to complete the synchronization update is calculated. T delay The specific calculation formula is as follows: in, t 0 represents the update initiation time. t 1, t 2, t 3, ... represent the time when the provincial, municipal, and county-level nodes completed the synchronization update, respectively; Generate a version number that includes the update initiation time, geographic entity status, and operator metadata. V The specific generation method is as follows: in, e For identifying geographical entities that have undergone significant changes, E rel For the identification of a set of associated geographic entities, It is the hyperbolic tangent function.

[0015] Preferably, in step 3, the event-driven model specifically includes an event listener, an event resolver, and an update scheduler, wherein: The event listener monitors for significant changes to geographic entities in real time and encapsulates the events into standard event objects containing geographic entity identifiers, change types, and timestamps. The event parser receives standard event objects, extracts geographic entity identifiers from them, and performs a breadth-first search based on the business-spatial-temporal knowledge graph to identify the set of associated geographic entities. At the same time, it parses the change type to determine the type of update operation to be triggered. The update scheduler calculates the update priority of each associated entity based on the set of associated geographic entities output by the event resolver and the update operation type, combined with the edge attributes in the knowledge graph. It then generates an update task queue according to the priority order and schedules and executes the update tasks in the queue in an asynchronous and non-blocking manner.

[0016] As a preferred approach, in the process of using relative entropy to assess the consistency of data distribution before and after the update, a comprehensive consistency assessment index is also generated, which specifically includes: In calculating relative entropy Simultaneously, spatial distribution features and attribute statistical features are extracted based on the spatial geometric shape and attribute value set of geographic entities, respectively. Spatial distribution features include the logarithm of entity area, the logarithm of entity perimeter, and entity shape index; attribute statistical features include the mean, variance, and coefficient of variation of all attribute values. The extracted spatial distribution features and attribute statistical features are combined in a preset order to form a multi-dimensional feature vector representing the state of geographic entity data. The mean vectors of all geographic entity corresponding feature vectors in the data distribution P before the update and the data distribution Q after the update are calculated respectively. The Euclidean distance between the two mean vectors is calculated, and this Euclidean distance is compared with the relative entropy. We perform weighted fusion to obtain a comprehensive consistency evaluation index.

[0017] The present invention has at least the following beneficial effects: (1) By constructing a unified knowledge graph, business rules, spatial and temporal relationships are formalized, and a closed loop from change detection, association discovery to triggering updates is realized, which significantly reduces manual judgment and intervention, and enables cross-departmental and cross-business data collaborative updates to be executed efficiently and automatically; (2) By adopting a standard knowledge representation language to formalize business rules and quantitatively modeling spatial adjacency and temporal correlation, implicit business logic and fuzzy natural relationships are transformed into precise, machine-executable and computable structures, providing a foundation for efficient, accurate and intelligent processing; (3) By combining hash comparison and spatial geometric difference operation for refined change detection, and using a quantitative model to filter insignificant changes; then, by using knowledge graph multi-path search to accurately locate related entities, and combining multiple (3) Dimensional factors are used to calculate update priority, and asynchronous and orderly scheduling is carried out through event-driven architecture to ensure the accuracy of update triggering and the efficiency of system processing; (4) Conflicts are detected by quantifying spatial overlap and automatically resolved based on objective business priorities, which effectively solves the contradiction problem in concurrent updates. Statistical methods such as relative entropy are introduced to verify the consistency of data distribution before and after the update, which can discover unexpected deviations at the macro level and enhance the reliability of update results; (5) Update timeliness is guaranteed by monitoring cross-level synchronization delay. Full lifecycle management of each linkage update is realized by generating version numbers containing information and recording complete metadata. By constructing comprehensive consistency evaluation indicators, multi-dimensional and more robust data quality evaluation methods are provided, which significantly improves the transparency and controllability of the entire data governance process. Attached Figure Description

[0018] Figure 1 A schematic diagram illustrating a technical verification of the collaborative update method provided by the present invention; Figure 2 This is a schematic diagram of an influence domain for the geographic entity linkage update of the present invention. Detailed Implementation

[0019] The present invention will now be described in further detail with reference to the accompanying drawings, so that those skilled in the art can implement it based on the description.

[0020] like Figure 1 As shown, the present invention provides a knowledge-driven spatiotemporal data collaborative update method, which includes the following steps: Step 1: Analyze the natural resource business process and extract business rules. Use OWL2 language to convert the business rules into machine-executable rules. Calculate the spatial adjacency of geographic entities and the temporal correlation of geographic entity states based on the standard definition of topological relationships. With geographic entities as core nodes, and machine-executable rules, adjacency, and temporal correlation as edge attributes, construct a business-space-temporal knowledge graph.

[0021] The construction of a business-space-temporal multidimensional knowledge model establishes a unified computational model for natural resource data that integrates business logic, spatial relationships, and temporal evolution. This transforms previously scattered and implicit business rules, spatial constraints, and temporal dependencies into structured knowledge that is computer-recognizable and computable. Specifically, this first requires in-depth analysis of the entire business process in the natural resource field, including "measurement, adjustment, planning, approval, supply, use, replenishment, inquiry, and registration," to extract key business rules and constraints. These rules typically exist in natural language or internal regulations. To enable automatic execution by the system, they need to be formalized using a knowledge representation language such as OWL2, converting them into explicit "if-then" logical statements or machine-executable rule sets. For example, a rule trigger function can be defined; when a specific attribute of an approval event (such as approval status) falls into a preset threshold set (such as "approval passed"), the rule is considered activated. Simultaneously, it is necessary to quantitatively model the spatial relationships between geographic entities. For example, by calculating the length ratio of the shared portion of the boundary between two plots, their adjacency can be defined, thus transforming the vague spatial concept of "adjacency" into a comparable numerical indicator, such as setting a ratio exceeding 0.1 to determine adjacency. In the temporal dimension, a time decay function is introduced to quantify the correlation strength of entity states at different points in time; the longer the time interval, the weaker the correlation. Finally, each specific geographic entity (such as a piece of land or a construction project) serves as a core node in the knowledge graph. The formalized business rules, the calculated spatial adjacency values, and the temporal correlation values ​​are used as "edges" or "edge attributes" connecting different entity nodes, collectively weaving together a three-in-one knowledge graph of "business-space-temporal." This graph not only stores the static attributes of entities but also dynamically depicts the complex linkage relationships between entities caused by business, space, and time, providing a data and relational foundation for subsequent intelligent sensing and linkage updates.

[0022] Step 2: Based on log analysis, time series comparison and spatial overlay technology, extract the attribute change increment of structured business data and the spatial change increment of spatial data, and calculate the comprehensive change degree of geographic entities based on the attribute change increment and spatial change increment. When the comprehensive change degree exceeds the preset threshold, it is determined that the geographic entity has changed significantly.

[0023] Intelligent detection of incremental changes in multi-source data automatically and accurately identifies geographical entities that have undergone substantial changes from massive, dynamic natural resource data. Due to the diverse data sources, including structured business databases (such as approval forms) and unstructured spatial databases (such as vector maps), this method employs a combined incremental detection strategy. For structured business data, changes in the hash values ​​of key fields are efficiently used to determine whether attributes have changed. For example, if the hash value of the "user" field for a land parcel is found to be inconsistent, an update to that attribute can be quickly identified. For spatial data, a geometric calculation method is used to accurately identify newly added, reduced, or deformed areas by comparing the symmetrical difference between old and new vector data. After obtaining the incremental changes in attributes and space, not all subtle changes require triggering a linkage mechanism; therefore, a comprehensive change degree index is defined to assess the significance of the changes. This index combines the proportion of spatial change (such as the proportion of changed area to original area) with the breadth and importance of attribute changes (considering the preset weights of different attributes) through a weighted formula, forming an evaluation value between 0 and 1. For example, it can be set that a significant change has occurred in a geographic entity only when the evaluation value exceeds 0.05 (this threshold can be adjusted according to the actual business sensitivity; options such as 0.01, 0.03, or 0.1 are all possible). This process transforms raw data into a change event, providing a precise trigger signal for linked updates.

[0024] Step 3: When a geographic entity undergoes significant changes, a breadth-first search is performed based on the business-spatial-temporal knowledge graph to identify the set of associated geographic entities related to the geographic entity, and an event-driven model is used to trigger the synchronous update of the set of associated geographic entities.

[0025] When a geographic entity is determined to have undergone significant changes, the system does not update that entity in isolation. Instead, based on the constructed knowledge graph, it intelligently discovers and updates all related entities, achieving a chain reaction of updates. The discovery of related entities is primarily accomplished through a breadth-first search on the knowledge graph, based on various relationships encoded in the graph's edge attributes: spatial adjacency (e.g., adjacency exceeding a threshold), strong temporal correlation (e.g., high correlation between recent states), and business rule correlation (e.g., according to an approval rule, a change in the status of plot A requires simultaneous verification of plot B). The set of related entities obtained from the search constitutes the target scope requiring simultaneous updates. To optimize the update order and resource allocation, the system calculates an update priority for each related entity in the set. This priority comprehensively considers the entity's spatial proximity to the source entity that has changed, its temporal correlation, and its business logic dependency, derived through a weighted fusion formula. Subsequently, the system employs an event-driven model to schedule update tasks: the change of the source entity is encapsulated as a standard update event, which triggers asynchronous, non-blocking update operations on each related entity according to the calculated priority order. For example, the update scheduler will generate a task queue to prioritize updates of spatially adjacent and business-urgent related entities, ensuring that critical linkages can be completed as soon as possible.

[0026] Step 4: Calculate the spatial overlap between associated geographic entities during the synchronous update process. When the spatial overlap exceeds the preset overlap threshold, it is determined to be a spatial conflict. The spatial conflict is resolved based on business priority, and the consistency of data distribution before and after the update is evaluated using relative entropy.

[0027] When updating multiple associated entities concurrently or in a chain, conflicts between update results may occur. The most common one is spatial conflict: for example, two different update events attempt to assign mutually exclusive attribute states to the same spatial region. To address this issue, a conflict detection mechanism is introduced, which quantifies the severity of a conflict by calculating the spatial overlap degree between entities involved in the update (such as the ratio of the overlapping area to the area of the smaller one). For example, when this ratio exceeds 0.3 (the threshold is adjustable, such as 0.2, 0.25, 0.4, etc.), it is determined that there is a spatial conflict that needs to be resolved. Conflict resolution is not a simple crude selection between alternatives, but an intelligent adjudication based on business logic. The system calculates a business priority for each conflict-involved entity, which is comprehensively evaluated based on the importance level of the business type associated with the entity (for example, the level of land security-related businesses is higher than that of general management businesses) and its corresponding weight. When a conflict occurs, the update result required by the entity with higher business priority will overwrite the result of the entity with lower priority, so as to ensure the implementation of core business logic. After completing conflict resolution and update operations, consistency verification on the overall update result is also required to ensure that the logical integrity of the data has not been damaged. The verification method can adopt tools such as relative entropy (KL divergence) in information theory, and evaluates the consistency level by comparing the probability distribution difference of key attributes in the entire relevant data set before and after the update. If the difference is below a certain acceptable threshold, the linked update is considered to successfully maintain consistency.

[0028] Step 5: Calculate the time delay for regional nodes at all levels to complete synchronous update, and generate a version number containing metadata of the update initiation time, geographic entity status and operator to support version backtracking.

[0029] Natural resource data management often involves multi-level nodes such as national, provincial, municipal and county-level nodes. Ensuring that updates can be synchronized to all levels of nodes efficiently and consistently is crucial. After completing the core linked update, the method distributes the update results to all levels of nodes through standardized data interfaces and services. To monitor synchronization efficiency, the system calculates the maximum time delay from the moment the update is initiated to the moment each level of nodes confirms completion, and uses this as an indicator to evaluate the timeliness of the system. For example, the delay is required to be controlled within 60 seconds. Meanwhile, to meet the requirements of data audit and backtracking, each complete linked update process is recorded as a data version. The system automatically generates a unique version number that integrates information such as update time and core entity identifier. In addition, the version snapshot exhaustively records metadata such as the status of relevant entities before and after update, the operator who performed the update, and update timestamp. This version management system enables any historical update to be traceable and verifiable, and provides support for data quality management, responsibility identification and status rollback.

[0030] This technical solution, through the closed-loop connection of the above five steps, constructs a complete technical system encompassing knowledge modeling, change perception, intelligent linkage, conflict resolution, synchronous verification, and version traceability. Compared to existing technologies, this method effectively overcomes the data silos and update delays caused by heterogeneous data models, implicit update rules, and poor cross-departmental collaboration in the natural resources field. It significantly improves the automation and intelligence level of multi-source heterogeneous spatiotemporal data linkage updates, reducing the cost and error rate of manual intervention; enhances the accuracy and business relevance of linkage updates through knowledge-driven precise association and priority scheduling; ensures data logical consistency and integrity during large-scale concurrent updates using conflict resolution and consistency verification mechanisms; and achieves controllability, manageability, and traceability of the data update process through efficient cross-level synchronization and comprehensive version management, thereby significantly optimizing the overall efficiency and collaborative capabilities of natural resources data governance.

[0031] In another technical solution, in step 1, the business rules are converted into machine-executable rules using the OWL2 language, and the quantization function used for the rule triggering conditions is defined as follows: ; in, a k ( expected value ) is a business event expected value The k One attribute, V k This is the set of trigger thresholds corresponding to this attribute. m The total number of attributes associated with the rule; quantified. This is a Boolean value; when its value is 1, it indicates that the rule has been triggered.

[0032] The process of transforming business rules into a machine-executable form is not a simple translation of statements. It requires converting the implicit logic and constraints described in natural language and embedded in business processes such as "measurement, adjustment, regulation, approval, supply, use, replenishment, query, and registration" of natural resources into explicit structures that computers can directly parse and reason about. Using OWL2 (Web Ontology Language 2), a standardized knowledge representation language, is the core technical means to achieve this transformation. Based on descriptive logic, OWL2 allows for the definition of rich classes, attributes, and their logical relationships (such as equivalence, inclusion, and disjointness), thereby accurately characterizing the ontology concepts and rules of the business domain. For example, "Construction land approval" can be defined as an event class with data attributes such as "applicant unit," "plot number," and "approval status." A business rule, such as "when the approval status of a construction land application changes to 'approved,' its planning condition review process must be triggered," can be formally expressed using OWL2 axioms. This axiom associates the "approval status" attribute value with a specific triggering condition ("approved") and specifies the subsequent actions or rule chains that trigger it. This process transforms business process knowledge, which originally relied on manual interpretation and execution, into digital assets stored in a knowledge base that can be automatically invoked by the rule engine.

[0033] The quantitative judgment logic of rule triggering conditions is expressed through a rigorous mathematical function, jointly judging multiple key attribute values ​​of a business event. Each business event (e.g., a specific planning permit approval operation) contains several attributes, such as event type, initiating department, associated land parcel, approval conclusion, timestamp, etc. In the formal rule, a specific set of attributes of interest is set for the rule, and a set of trigger thresholds is defined for each attribute. This set of thresholds can be explicit state enumeration values, for example, the trigger threshold set for the approval conclusion attribute is (passed), "agree in principle"; or it can be a numerical range (e.g., the investment amount attribute must be greater than a certain threshold value). The trigger function is essentially a logical "AND" operation: the rule is determined to be in a triggered state (function value is true, or quantified as 1) only when all the attribute values ​​of the business event that the rule is interested in fall within their respective preset threshold sets. This design ensures the accuracy and controllability of rule triggering. Only when the event fully meets the multi-dimensional composite conditions preset by the rule will the corresponding linkage update logic be activated, avoiding false triggering caused by the satisfaction of a single condition.

[0034] During system initialization or rule updates, business experts or system administrators use specialized tools to model the refined business rules using the OWL2 language and import them into the rule knowledge base. When a new event is generated in the business system (such as a new approval record being generated in the database), the event listener component captures the event and standardizes it into an event object containing all relevant attributes. Subsequently, the rule engine loads the relevant formal rules and calls the trigger function to evaluate the event object. The engine iterates through the event's attributes and compares them with the set of attribute thresholds defined in the rule. For example, for a rule about "post-land supply supervision," it might focus on whether the "business type" attribute value is "land transfer," whether the "transfer fee" attribute is "fully paid," and whether the "contract signing" attribute status is "completed." The supervision rule will only be triggered when a land transfer event simultaneously meets all three conditions. After triggering, the rule engine outputs a trigger signal and the corresponding rule ID, which becomes the starting point for driving subsequent steps (such as related entity lookup and update task generation). The size of the attribute set (m-value) and the threshold set are entirely dependent on the complexity and granularity requirements of the business rules, and can be flexibly configured and adjusted in actual deployment. By formalizing the business rules using a standard knowledge representation language and defining a quantitative trigger function based on multi-attribute joint judgment, ambiguous business requirements are transformed into clear and computable logical instructions. This improves the depth of embedding business rules into the information system and enhances machine understandability, enabling precise and automatic execution of cross-departmental and cross-business linkage logic, significantly reducing the risks of ambiguity, delays, and errors caused by manual rule transmission and interpretation in traditional methods.

[0035] In step 1, the specific method for calculating the spatial adjacency of geographic entities is as follows: Based on the standard-defined topological relationships, calculate the spatial entities in two geographic entities. S 1 and S The adjacency degree of 2 is calculated using the following formula: in, Representing spatial entities S The boundary, Indicates the boundary length; when adjacent degree When they are, they are determined to be adjacent geographical entities; The specific method for calculating the temporal correlation degree of geographic entity status is as follows: Calculate geographic entities e At the point of time t 1 and t 2. State correlation The calculation formula is: in, λ This is the time decay coefficient, with a value ranging from 0.01 to 0.1; When constructing a business-spatial-temporal knowledge graph, geographic entities are used as core nodes, and machine-executable rules are incorporated. Adjacency and time correlation As edge attributes connecting geographic entity nodes, they are used in fusion modeling to obtain a unified representation of entities, relationships, and rules.

[0036] In natural resource management, "adjacency" is a crucial spatial topological relationship that directly impacts numerous operations such as ownership boundary confirmation, planning conflict detection, and ecological corridor continuity assessment. However, the simple "spatial intersection" judgment in traditional GIS is often too coarse, failing to distinguish between the two distinct degrees of adjacency: "slight contact" and "long boundary sharing." Adjacency calculation models use a continuous numerical value to finely characterize this degree of closeness. The calculation is based on the boundary geometry of two geographic entities (typically represented as polygonal features on a map, such as land parcels or administrative regions). Specifically, spatial operations are first used to obtain the shared boundary portion (i.e., boundary intersection) of the two entities, and then the length of this shared boundary is calculated. To ensure comparability, the length is divided by the smaller of the original boundary lengths of the two entities. This ratio is the adjacency degree, ranging from 0 to 1. When two entities are completely non-contacting, the adjacency degree is 0; when they share a partial boundary, the adjacency degree is a decimal greater than 0; theoretically, if the entire boundary of one entity completely overlaps with a portion of the longer boundary of another entity, its adjacency degree may approach 1. In practical applications, an empirical threshold can be set to determine whether a business-meaning "adjacent" relationship exists. For example, this threshold might be set between 0.05 and 0.2, with 0.1 being a common starting value. This means that the system considers two entities to be business-adjacent only when the shared boundary length reaches a certain proportion (e.g., one-tenth) of the smaller entity's own boundary length, potentially triggering linked updates based on the adjacency relationship (such as usage compatibility checks for adjacent plots). This quantification method provides richer information than traditional Boolean judgments, supporting priority ranking or impact analysis based on "adjacency strength."

[0037] Natural resource data exhibits strong temporal characteristics, with the state of entities (such as land use type, owner, and approval stage) constantly evolving over time. Understanding the strength of the correlation between states at different points in time is crucial for assessing the impact of historical states on the present and tracing change chains. A model based on an exponential decay function is used to quantify temporal correlation. The closer an entity's current state is to a past state, the stronger its influence or correlation with that historical state should be; this correlation gradually weakens over time. The key parameter in the model is the time decay coefficient, which controls the rate at which the correlation decreases with increasing time intervals. This coefficient is typically a small positive number, for example, ranging from 0.005 to 0.1. A smaller value means a slower decay, indicating that historical states are still considered relevant to the present over a longer time window; a larger value indicates a faster decay, with the system focusing more on recent changes. In specific calculations, the information of the entity at two time points (t1 and t2) is input, the absolute value of their time difference is calculated, and the correlation value between 0 and 1 is obtained by substituting this value into the exponential decay formula. For example, when the time decay coefficient is set to 0.05, the temporal correlation between two states with a 14-day interval (approximately two weeks) is about 0.5; after a 30-day interval, it drops to about 0.22. This provides a quantitative basis for judging the temporal continuity of entity states and identifying update events that may be related due to their proximity in time.

[0038] The knowledge graph is constructed with each independent geographic entity (assigned a unique ID) as a core node. The node's attribute table stores the entity's static and dynamic business attributes (such as area, usage, owner, current status, etc.). The rich and diverse relationships between entities are represented and carried by "edges." Edges not only identify the relationship type (such as "adjacent to," "occurred after," "constrained by rule"), but also store the calculated quantitative values ​​as attributes of the edge. For example, between plot A and plot B, an edge of type "spatial adjacency" can be created, recording the specific adjacency degree calculated using the method in Part One (e.g., 0.15). Similarly, between the "Y1 year status" node and the "Y1+1 year status" node of plot C, a "temporal association" edge can be established, recording the temporal association degree calculated by the model in Part Two (e.g., 0.8). Formalized business rules themselves can also be used as a special type of edge or as attributes of rule nodes, connecting with related entity nodes. This approach establishes a multi-dimensional, fused knowledge graph: nodes carry core entity information, and edges not only define relationship categories but also meticulously characterize the strength, closeness, and triggering conditions of relationships through quantified adjacency, correlation, and rule logic. This enables subsequent association analysis, influence reasoning, and update scheduling to perform more precise and intelligent retrieval and calculation from this quantitatively rich graph. For example, it prioritizes influencing entities with high spatial adjacency and tight temporal correlation, thus achieving a leap from binary judgments of "whether there is a relationship" to gradient decision-making based on "the strength of the relationship." By introducing precise mathematical models to quantify spatial adjacency and temporal correlation, and using these quantified values ​​as the core attributes of the knowledge graph, a refined and numerical representation of complex relationships between geographic entities is achieved. This significantly improves the system's ability to understand and process continuous and gradual relationships in the real world, making the triggering conditions for linked updates more scientific and reasonable, and the discovery of associated entities more accurate and comprehensive, thereby effectively avoiding the update omissions or over-linkages caused by coarse relationship judgments in traditional methods.

[0039] In another technical solution, step 2, the specific method for extracting the attribute change increments of structured business data is as follows: For structured business data, compare the hash values ​​of new and old fields to detect attribute changes. At that time, determine the attribute increment. ,in For hash functions, a new For new field data, a old This is old field data; The method for extracting spatial change increments from spatial data is as follows: Calculate two periods of spatial data Snew and S old The symmetric difference between them yields the spatial variation increment. ΔS ,Right now: ,in Added data This is a reduction in the amount of data. Calculate the overall degree of change of geographic entities C ( e The specific formula is as follows: in, α This represents the spatial weight, with a value ranging from 0.6 to 0.8. Represents the area of ​​a space; indicator function In attribute increment The value is 1 if the condition is met, and 0 otherwise. w k For the first k Preset weights for each attribute, n Total number of attributes; The preset threshold for the overall degree of change is θ ,when At that time, determine the geographic entity e Significant changes have occurred.

[0040] Structured business data is typically stored in relational database tables, such as project approval forms and land registration forms, with each record containing multiple fields. Traditional full-data comparison methods are inefficient when dealing with large datasets. This method uses hash functions as a fast filter for change detection. Hash functions can map data of arbitrary length to a fixed-length, nearly unique hash value (such as MD5 or SHA-256 algorithms). For key business fields that need to be monitored, the system periodically or in real-time calculates the hash value of its current value and compares it with the previously stored baseline hash value. If the hash values ​​are equal, it can be concluded with a high probability that the field content has not changed; if the hash values ​​are not equal, the field content must have changed, triggering precise difference calculation to determine the specific incremental content. For example, in land supply business, if the "Contract Signing Status" field changes from "Not Signed" to "Signed," the hash comparison will immediately detect the change and record the incremental information "Status Change = Signed." This process enables lightweight and real-time perception of changes in massive amounts of business data, avoiding the need for word-by-word content analysis of all data, and concentrating computing resources on the small amount of data that has actually changed.

[0041] Changes in spatial data encompass not only attributes but, more importantly, alterations in geometry and location. This method draws upon the concept of set operations in computational geometry, identifying spatial increments by calculating the symmetric difference between two spatial data layers from different periods. In practice, the system invokes the geometric difference operation function of a spatial database (such as PostGIS) or a GIS computing engine. For two versions of spatial graphics of the same geographic entity or the same area (e.g., plot A in version Y2 and plot A in version Y2+1), the symmetric difference operation accurately outputs two parts: one part represents the graphic area that exists only in the new version but not in the old version (i.e., the newly added portion), and the other part represents the graphic area that exists only in the old version but not in the new version (i.e., the reduced or lost portion). The union of these two parts constitutes the complete spatial change increment ΔS. For example, if a plot expands outward by a corner due to planning adjustments, the newly added portion is the polygon of that corner; if another plot is divided, the reduced portion of the original plot is the polygon that was divided out. This method can clearly and comprehensively capture any form of spatial geometric change, whether it is expansion, contraction, segmentation, or merging, providing a precise geometric description for understanding spatial evolution. However, simply perceiving change is insufficient, as minor attribute modifications or boundary fine-tuning may not warrant large-scale, coordinated updates. Therefore, a comprehensive change index C(e) is needed to assess the operational importance of change. This index considers both the proportion of spatial change and the breadth and weight of attribute changes. The contribution of spatial change is measured by calculating the proportion of the changed area to the original entity area, reflecting the magnitude of the physical change. The contribution of attribute changes is calculated by counting the number of changed attribute items and considering the pre-defined operational weights of each attribute (e.g., the weight of "land use" is much higher than that of "remarks"). Finally, an adjustable spatial weighting factor α (e.g., ranging from 0.5 to 0.8, with 0.7 being a common reference value) combines these two parts. A larger α value indicates a more dominant role of spatial change in the comprehensive assessment. The calculated C(e) value is a number between 0 and 1. The system presets a significance threshold θ (for example, it can be set to 0.03, 0.05, or 0.1, adjusted according to business tolerance and update costs). Only when C(e) ≥ θ does the system ultimately determine that the geographic entity has undergone a significant change, thus qualifying it to trigger the subsequent linked update process. This mechanism ensures that the system can intelligently distinguish between noise and signal, and only when the change reaches a certain business significance level will the costly global linkage be initiated, effectively avoiding system instability and resource waste caused by frequent and trivial changes while ensuring update sensitivity.

[0042] By combining hash comparison with geometric difference operations, a refined and efficient capture of changes in multi-source heterogeneous data is achieved. Furthermore, by introducing a comprehensive change assessment model with configurable weights, an intelligent filtering and decision-making mechanism is established, transforming changes in raw data into events of business significance. This significantly improves the automation and accuracy of change detection, ensuring that the linked update process is triggered only when truly necessary, thereby optimizing overall system performance and resource utilization, and reducing unnecessary computational overhead and business interference.

[0043] In another technical solution, step 3 involves performing a breadth-first search based on a business-spatial-temporal knowledge graph to identify the set of associated geographic entities related to the geographic entity. E rel The specific identification and judgment conditions are as follows: in, expected value To act on geographic entities e And business events that triggered significant changes in it, e' Indicates candidate associated geographic entities, conditions , , These correspond to spatial adjacency, temporal association, and business rule association between two geographical entities, respectively. Synchronous updates are triggered using an event-driven model, specifically including: For the set of associated geographic entities E rel Each geographic entity in e' Calculate update priority P ( e' The calculation formula is: in, β and γ The weighting coefficients and β+γ≤1 , The association weights are assigned based on business rules; Based on the update priority of each geographic entity, the synchronous update of the associated geographic entities is triggered and executed.

[0044] Once a geographic entity e is determined to have undergone significant changes, a breadth-first search is performed on a pre-constructed "business-space-temporal" knowledge graph, starting from e, to intelligently locate the affected set of related entities E. relBreadth-first search ensures that entities directly related to `e` are discovered first, and then expanded layer by layer, balancing efficiency with the hierarchical nature of the relationships. The search selection criteria are based on the quantified relationship values ​​stored on the graph edge attributes, specifically including three optional association channels: first, the spatial adjacency channel, requiring that the spatial adjacency between candidate entity `e'` and `e` is not lower than a certain threshold (such as 0.1 mentioned above, which can actually be fine-tuned between 0.08 and 0.15); second, the temporal association channel, requiring that the state association between `e'` and `e` within a specific time window is not lower than another threshold (such as 0.5, meaning the association needs to reach a moderate strength or higher, this threshold can be explored between 0.4 and 0.6); and third, the business rule channel, where, based on the formal rule base, when the business event `ev` applied to `e` is also related to `e'` (i.e., R...), the search criteria are used to determine when the rule triggering condition is also related to `e'`. trigger If (ev, e') is true, then e' is included. If any one of these three conditions is met, e' is considered a related entity. For example, a change in the planned use of a plot of land (event ev) not only directly affects itself (e), but also triggers compatibility checks on adjacent plots (meeting spatial conditions) based on business rules. Simultaneously, based on temporal correlation, it affects the status synchronization of another plot with which it recently had a transfer of ownership. Through this multi-path correlation discovery, the system can construct an update impact network centered on the source of change, covering multi-dimensional impacts in business, space, and time.

[0045] The priority calculation model integrates the correlation strength between the related entity e' and the source entity e across different dimensions. Higher spatial adjacency indicates a more direct and urgent spatial impact, thus occupying a portion of the priority (β). Higher temporal correlation indicates a tighter coupling between the two states on the timeline, strengthening the necessity for synchronous updates, and occupying another portion of the weight (γ). Furthermore, the business rules themselves define the importance of different relationships, which is determined by an independent business rule weight R. weight The priority P(e, e') is represented by a predefined urgency label or impact level in the rule base. β and γ are pre-set coefficients that satisfy β + γ ≤ 1 to ensure the rationality of weight allocation; for example, β = 0.4 and γ = 0.3. The priority P(e') is a numerical value; the higher the value, the greater the urgency for the related entity to be updated. This quantitative ranking mechanism ensures that critical and urgent linkages are responded to first. Figure 2 The update example shown illustrates how, after a geographic entity k0 is determined to have undergone a significant change, the search locates the affected set of related entities E. relThis includes geographic entities k1, k2, k3, and k4. This is the influence domain of the data that needs to be updated when geographic entity k0 undergoes a significant change. The relational values ​​of 0.46, 0.57, 0.97, and 0.62 correspond to the update priorities of geographic entities k1, k2, k3, and k4. Updates to related geographic entities are performed according to the order of these update priority values.

[0046] The event-driven, linked update execution process consists of several cooperating components: An event listener continuously monitors the data layer; once it detects a significant change in a geographic entity, it encapsulates it into a standard message containing information such as the event source, change type, and timestamp. The event parser receives this message, first extracting the identifier of the changed entity *e*, and then drives the knowledge graph query engine to perform a breadth-first search using the conditions described in Part I to find the set of all related entities *E*. rel Simultaneously, the change type is analyzed to determine the specific update operation to be performed (such as state update, area recalculation, notification generation, etc.). Finally, the update scheduler receives the parsed results and calls the priority model E. rel Each entity in the algorithm calculates P(e'), and then, based on priority from high to low, these entities and their corresponding update operations are packaged into independent update tasks and placed in a priority queue. The scheduler retrieves tasks from the queue asynchronously and non-blockingly and distributes them to the corresponding data services or business systems for execution. This architecture decouples and pipelines triggering, analysis, scheduling, and execution, enabling the system to handle high-concurrency update events with ease, prioritizing the spread of core impacts while maintaining the overall system's high efficiency and stability. A multi-condition composite entity discovery mechanism ensures the comprehensiveness and accuracy of update impact scope identification; a multi-factor fusion update priority quantification model enables intelligent sorting and resource optimization for linked update task execution; and an event-driven architecture and clear role division construct an efficient, reliable, and scalable linked update execution engine. This significantly improves the organization, timeliness, and controllability of large-scale, cross-domain data linked updates, effectively avoiding problems such as chaos, delays, or the overloading of important updates during the update process.

[0047] In another technical solution, step 4 uses the following formula to calculate geographic entities. e 1 and e 2 Between: When spatial overlap At that time, it was determined to be a spatial conflict; The process of resolving spatial conflicts based on business priorities specifically includes: Calculate business priorities for conflicting geographic entities P bus ( e The calculation formula is: in, b k Representing geographic entities e The associated first k Item business type, level ( b k This indicates the preset level for this service type. w k The preset weight for this business type, m This represents the total number of associated business types. When conflicts are resolved, the update results of geographic entities with higher business priority will override the update results of geographic entities with lower business priority.

[0048] Spatial conflicts typically occur when update operations on two or more geographic entities attempt to modify the same physical spatial area, assigning it contradictory or mutually exclusive attribute states. To objectively measure the severity of the conflict, a quantitative metric, spatial overlap, is introduced. This involves calculating the geometrical intersection area of ​​the two entities involved in the update dispute (e.g., updated plot A and updated plot B) and comparing this intersection area to the area of ​​the smaller of the two entities, obtaining a ratio. This ratio directly reflects the spatial proportion or scope of the conflict. For example, if the two updated plots completely overlap, the overlap will reach its maximum value; if they only slightly intersect at the edges, the overlap will be a smaller value. To determine whether a conflict warrants intervention, the system presets an overlap threshold. This threshold is an empirical value used to distinguish between accidental, negligible minor overlaps and substantial, potentially contradictory, overlaps. For example, this threshold can be set between 0.2 and 0.4, with 0.3 being a commonly used starting point. This means that the system will only formally determine a spatial conflict and submit it to the resolution module for processing when the calculated spatial overlap reaches or exceeds 30% (i.e., more than one-third of the area of ​​the smaller entity overlaps with the updated range of another entity). This quantification threshold mechanism effectively filters out a large number of non-substantive geometric contacts, focusing processing resources on conflict events that may actually cause data discrepancies and business chaos.

[0049] When spatial competition occurs, priority should be given to ensuring the achievement of more important and core business objectives. To this end, the system needs to calculate a quantified priority score for each geographic entity involved in the conflict. This score is calculated based on all business types associated with that entity. In natural resource management, different businesses have different strategic importance and urgency levels. For example, businesses involving adjustments to ecological protection red lines are typically far more important than general land registration information corrections; businesses involving the implementation of major construction projects are more urgent than routine planning and maintenance. The system pre-defines an importance level for each type of business (e.g., divided into levels 1 to 5) and can assign different weights based on specific scenarios. When calculating the business priority of an entity, the system iterates through all business matters currently associated with that entity, multiplies their respective importance levels by their corresponding weights, and sums them up to obtain a comprehensive score. This score quantifies the total business value represented by that entity in this conflict. For example, a plot of land simultaneously associated with an ecological restoration project (high level) and temporary land use approval (low level) will have a significantly higher business priority score than a plot only associated with routine cadastral surveys. Through this calculation, conflict resolution is transformed from a vague, subjective judgment into a traceable, objective decision-making process based on established business rules.

[0050] When two entities e1 and e2 are determined to have a spatial conflict, and their respective business priorities P bus (e1) and P bus Once (e2) is calculated, the resolution logic is clear and definite: the update result corresponding to the entity with higher business priority overwrites the update result of the entity with lower priority. For example, if P bus (e1)>P busIf (e2) is requested, then the database will ultimately store the updated spatial extent and attribute status of entity e1. Update requests for entity e2 will be rejected or corrected due to spatial conflict and insufficient priority (e.g., a rollback to the pre-conflict state or a system-triggered coordination process requiring manual intervention). The entire resolution process is automatically executed by the system's conflict resolution module. Upon receiving an alert from the conflict detection module, this module immediately queries the business association information of the conflicting entities, calls the priority calculation model, compares the results, and automatically sends a final ruling instruction to the data update engine. Simultaneously, the complete conflict event, involved entities, priority calculation basis, and resolution results are all recorded in detail in the log for auditing and traceability. This ensures that, while safeguarding core business objectives, contradictions in data updates are automatically resolved, maintaining the consistency of the data spatial logic. By introducing a quantifiable spatial overlap threshold detection, the system achieves accurate identification of substantive spatial conflicts. By establishing a priority calculation model based on business rules and importance levels, it provides an objective, fair, and business management-compliant decision-making basis for conflict resolution. Finally, through an automated adjudication and execution mechanism, it significantly improves the efficiency and intelligence of handling spatial data update conflicts, effectively avoiding data inconsistencies and business stagnation caused by slow manual coordination or subjective judgment differences.

[0051] In step 4, relative entropy is used to evaluate the consistency of the data distribution before and after the update, specifically including: Calculate the data distribution before the update P Compared with the updated data distribution Q The relative entropy between : in, N The total number of values ​​that the data can take; when If the value is less than the preset consistency threshold, the update is deemed to meet the consistency requirements.

[0052] After a series of linked updates, although changes to individual records may conform to business rules, from a macro-statistical perspective, whether the overall distribution characteristics of a large amount of data have undergone unexpected and unreasonable deviations is an important dimension for measuring update quality. Relative entropy, a core concept in information theory, measures the difference between two probability distributions. It is the average amount of information lost when we use the ideal updated distribution Q to approximate the true pre-update distribution P. If P and Q are exactly the same, the relative entropy is zero; the greater the difference, the greater the relative entropy value. In this method, P represents the value distribution of all relevant geographic entities on one or a set of key attributes before the linked update; Q represents the value distribution of the same batch of entities on the same attribute after all linked update operations are completed. Introducing relative entropy into consistency verification goes beyond simply checking whether each record was updated successfully. Instead, it examines from a more macro-statistical perspective whether this batch update has changed the inherent patterns and structure of the data as a whole. This helps to discover potential risks where, although each update conforms to local rules, the aggregate effect may lead to unreasonable and drastic changes in overall data characteristics (such as the proportion of various land uses, the proportion of different approval statuses, etc.).

[0053] The data distribution of geographic entities is defined as a probability distribution. For example, a key classification attribute, such as "land use type," can be chosen. Before the update, the area or number of plots of each land use type (e.g., cultivated land, construction land, forest land, etc.) is statistically analyzed across all affected plots, forming a discrete probability distribution P. Correspondingly, after the update, the same proportions are recalculated based on the latest data to obtain distribution Q. Then, according to the formula for calculating relative entropy, for each land use type i, P(i) is multiplied by the logarithm of the ratio of P(i) to Q(i), and the results for all types are summed to obtain the probability distribution Q. The value is calculated automatically by the system's consistency verification module. This module starts after the update scheduler has completed all tasks, extracts relevant data snapshots before and after the update from the database, and calculates the relative entropy value one by one according to the preset list of key attributes. In addition to categorical attributes, for continuous numerical attributes (such as land parcel area), probability distributions can also be constructed by discretizing them into multiple intervals (such as area segments) for calculation. This process achieves precise quantification of changes in the overall state of the data.

[0054] Calculate Next, a standard is needed to determine whether this difference is within an acceptable range. The system presets a consistency threshold, which is a small positive number representing the maximum distribution deviation that the business can tolerate. For example, this threshold can be set to 0.01, 0.05, or 0.1, depending on the business's requirements for data stability; 0.05 is a common reference starting point. The judgment logic is simple: if the calculated relative entropy value is less than the preset consistency threshold, the update is considered consistent at the macro-statistical level, meaning the update has not caused unacceptable distortion of the overall data characteristics, and the verification passes. Conversely, if the relative entropy value is greater than or equal to the threshold, it is judged as inconsistent, and the system will trigger an advanced alert. This alert indicates that the update may have unforeseen chain reactions or rule loopholes, causing a significant shift in the data distribution. The system can submit this result, along with a detailed distribution comparison report, to the system administrator or business expert for manual review. By introducing relative entropy from information theory as an evaluation tool, the connotation of data consistency is expanded from record-level correctness to the stability of macro-statistical distribution, providing a deeper dimension of quality measurement. Its automated calculation and threshold determination process can efficiently and objectively identify potential systemic deviation risks in batch updates, significantly enhancing the ability to control the overall quality of complex linkage update results.

[0055] In another technical solution, step 5 involves calculating the time delay for each level of nodes to complete the synchronization update. T delay The specific calculation formula is as follows: in, t 0 represents the update initiation time. t 1, t 2, t 3, ... represent the time when the provincial, municipal, and county-level nodes completed the synchronization update, respectively; Generate a version number that includes the update initiation time, geographic entity status, and operator metadata. V The specific generation method is as follows: in, e For identifying geographical entities that have undergone significant changes, E rel For the identification of a set of associated geographic entities, It is the hyperbolic tangent function.

[0056] In distributed or hierarchical natural resource data systems, there is typically a time lag between the initiation of an update (e.g., at a county-level node) and its eventual synchronization to nodes at all levels (city, province, and even national level). To ensure the eventual consistency of the "single map" data and meet business requirements for data freshness, the time consumption of this synchronization process must be quantitatively monitored. This method defines a synchronization delay calculation model that captures the time span from the moment the update command is issued at the source node to the moment the slowest node confirms completion. Specifically, the system assigns a globally unique timestamp t0 to each linked update transaction at the source node (e.g., the county-level system initiating the update). Subsequently, when update data packets or synchronization commands are distributed to various upper-level or peer nodes via the network, each node, upon successfully completing local data entry, verification, and returning a confirmation message, also attaches its own completion timestamp (e.g., t1 represents the completion time of the provincial node, t2 represents the completion time of the city-level node, and so on). Synchronization delay T delay The difference between the maximum completion time of all nodes requiring synchronization and the initiation time t0 is calculated. This calculation method, based on the slowest node, reflects the actual time taken for the update to reach a globally consistent state, conforming to the strict business definition of "all nodes in place." This metric provides a crucial basis for system performance evaluation and network optimization. In actual deployment, an acceptable latency limit can be set based on business urgency. For example, for non-urgent daily updates, this limit might be set at a few minutes; for critical business changes, it might require completion within one minute or even less. The system can monitor this metric in real time and issue an alert when the limit is exceeded.

[0057] In frequently updated data environments, version numbers are core identifiers for tracking data evolution history and supporting backtracking and comparison. This method's version number generation algorithm creates an identifier that is not only unique but also reflects the characteristics of the updated content to a certain extent, fusing the update's initiation timestamp with the core content identifier of this update. The timestamp t0, as the foundation of the version number, ensures the temporal order of versions. Simultaneously, it combines the unique identifier (e) of the source geographic entity that has undergone significant changes with the identifier (E) of the set of all associated geographic entities identified through the knowledge graph. relThe system performs some form of combination (e.g., string concatenation or hashing) and maps the result through a hyperbolic tangent function (tanh). The hyperbolic tangent function maps any real number input to a fixed interval (-1, 1). Its function is to standardize and compress potentially very large or irregular values ​​generated by combinations of entity identifiers, making them a stable, controllable small increment, which is then added to the timestamp. The resulting version number V contains information about the initiation time accurate to milliseconds or microseconds, and also embeds a digital fingerprint of the scope of the update's impact through the function mapping. This gives the version number semantic information, rather than simply a monotonically increasing sequence number. This helps in quickly linking the version number to a specific set of update events during subsequent auditing or troubleshooting.

[0058] Version numbers are generated for version management, and a complete data snapshot or change log is created for each version. In addition to the version number V and update initiation time t0, the system automatically captures and persistently stores a series of metadata crucial for auditing, backtracking, and problem diagnosis. These primarily include: 1) key state snapshots of all geographic entities involved before and after the update (including source and related entities), such as their spatial geometry and all business attribute values; 2) the user or system service identity (operator) who performed the linked update operation; 3) a list of business rule IDs triggered and applied throughout the update process; 4) conflict detection and resolution results records; and 5) synchronization completion timestamps for each level of node. This metadata is associated with the version number V and is structured and stored in a dedicated version management repository or log database. When it is necessary to backtrack to a historical version, the system can retrieve the complete metadata based on the version number and use this information to reconstruct the data state at that time, or clearly display all the details of changes from the previous version to the current version. This comprehensive version management mechanism transforms every data change into a traceable, auditable, and reproducible event, providing a solid technical foundation for data quality control, accountability, and emergency rollback. By accurately calculating the synchronization latency based on the slowest node, it provides objective monitoring indicators and optimization directions for cross-level data synchronization performance. Through the design of a version number generation algorithm that integrates time and content characteristics, it creates a unique version identifier that combines temporal sequence and a certain degree of semantic recognition. By systematically recording rich update metadata, it constructs a complete and traceable historical archive of data evolution. These measures significantly improve the transparency, controllability, and reliability of large-scale distributed data collaborative update processes, effectively ensuring that the status of data is knowable, its history is traceable, and problems are traceable during cross-regional and cross-level data flow, providing underlying support for the authority and credibility of natural resource data.

[0059] In another technical solution, step 3 employs an event-driven model that specifically includes an event listener, an event resolver, and an update scheduler, wherein: The event listener monitors for significant changes to geographic entities in real time and encapsulates the events into standard event objects containing geographic entity identifiers, change types, and timestamps. The event parser receives standard event objects, extracts geographic entity identifiers from them, and performs a breadth-first search based on the business-spatial-temporal knowledge graph to identify the set of associated geographic entities. At the same time, it parses the change type to determine the type of update operation to be triggered. The update scheduler calculates the update priority of each associated entity based on the set of associated geographic entities output by the event resolver and the update operation type, combined with the edge attributes in the knowledge graph. It then generates an update task queue according to the priority order and schedules and executes the update tasks in the queue in an asynchronous and non-blocking manner.

[0060] The event listener component is responsible for continuously and in real-time sensing "significant changes" occurring in the data layer and transforming these raw events into standardized messages that the system can process internally. Event listeners can be implemented using various technologies, such as database-based triggers, listening to database transaction logs (e.g., MySQL's binlog or PostgreSQL's WAL), subscribing to change notifications in message queues (e.g., Kafka, RabbitMQ), or directly interfacing with business system API calls. Regardless of the technology used, the goal is consistent: once a geographic entity is detected as having undergone a significant change, the event is immediately captured. After capture, the listener does not directly process complex business logic but performs initial encapsulation and standardization. It extracts the core elements of the event: which geographic entity changed (its unique identifier), what type of change occurred (e.g., "spatial geometry change," "ownership attribute change," "approval status change," etc.), and the exact timestamp of the event. This information is packaged into a structured standard event object, which follows a predefined protocol format to ensure unambiguous parsing by subsequent components. This process unifies change notifications from various heterogeneous underlying data sources into a high-level, abstract event flow.

[0061] The event resolver operates in steps. First, it extracts the identifier of the core geographic entity (i.e., the source of change, e) from the event object. Then, it accesses the pre-constructed "business-spatial-temporal" knowledge graph, using this entity as a starting point to execute a breadth-first search algorithm. Leveraging the quantified adjacency, temporal correlation, and formalized business rules stored on the graph's edge attributes, it intelligently derives the set E of all potentially affected related geographic entities. relThis step expands the change of a single entity into a potential "influence network." Simultaneously, the parser analyzes the type of change in the event object. For example, if the change type is "land use change," the parser, combined with rules from the knowledge graph, determines that a land use compatibility check operation needs to be triggered for the associated entities; if it's land parcel splitting, operations such as adjacent parcel boundary reconstruction and property area recalculation might be triggered. Finally, the event parser outputs a structured parsing result, which includes the source entity, the set of associated entities, and a list of specific "update operation types" to be performed for each associated entity or group of entities. After receiving the parsing result, the update scheduler updates the associated entity set E... rel For each member in the knowledge graph, its update execution priority is calculated. Priority calculation again utilizes information from the knowledge graph (such as specific values ​​for spatial adjacency and temporal correlation) and business rule weights, using a pre-defined formula to calculate a priority score for each entity to be updated. Based on this score, the scheduler sorts all tasks to be executed, forming an update task priority queue. Each task in the queue explicitly points to a target entity, one or more update operations, and the required parameters. Subsequently, the scheduler schedules these tasks asynchronously and non-blockingly. This means that the scheduler does not wait for one task to finish executing in the database before dispatching the next; instead, it dispatches the task to a dedicated data update executor or worker thread pool and immediately returns to continue processing the next task in the queue or wait for new events. The executor independently completes specific I / O-intensive operations such as database writes and service calls. This asynchronous and non-blocking architecture greatly improves the throughput and response speed of the entire system, preventing individual time-consuming update operations from blocking the entire event processing pipeline, thus ensuring that the system remains smooth and stable when facing high-concurrency update events.

[0062] This solution constructs a highly cohesive, loosely coupled, easily scalable, and maintainable collaborative update execution engine by explicitly decomposing the event-driven model into three clearly defined and interface-defined components: listening, parsing, and scheduling. The event listener achieves real-time and unified change awareness; the event parser, through deep utilization of knowledge graphs, endows the system with powerful intelligent reasoning capabilities, accurately defining the update impact domain; and the update scheduler, through priority sorting and asynchronous non-blocking scheduling, ensures the high efficiency of update task execution and the overall high concurrency processing capability of the system. This architecture significantly improves the intelligence level, processing efficiency, and robustness of the collaborative update system, enabling it to readily handle the complex, frequent, and demanding dynamic data collaboration needs in the natural resources field.

[0063] In another technical solution, during the process of using relative entropy to assess the consistency of data distribution before and after the update, a comprehensive consistency assessment index is also generated, which specifically includes: In calculating relative entropy Simultaneously, spatial distribution features and attribute statistical features are extracted based on the spatial geometric shape and attribute value set of geographic entities, respectively. Spatial distribution features include the logarithm of entity area, the logarithm of entity perimeter, and entity shape index; attribute statistical features include the mean, variance, and coefficient of variation of all attribute values. The extracted spatial distribution features and attribute statistical features are combined in a preset order to form a multi-dimensional feature vector representing the state of geographic entity data. The mean vectors of all geographic entity corresponding feature vectors in the data distribution P before the update and the data distribution Q after the update are calculated respectively. The Euclidean distance between the two mean vectors is calculated, and this Euclidean distance is compared with the relative entropy. We perform weighted fusion to obtain a comprehensive consistency evaluation index.

[0064] When assessing consistency, the system considers not only the probability distribution (relative entropy) of attribute values ​​but also actively extracts and integrates multi-dimensional features of geographic entities in terms of spatial geometry and attribute value sets to construct richer descriptive vectors. For spatial distribution features, several key shape indicators are automatically calculated from the vector geometry of each geographic entity, such as the logarithm of the entity area (reflecting scale), the logarithm of the entity perimeter (reflecting boundary complexity), and entity shape indices (e.g., perimeter-to-area ratio, used to quantify whether the shape is compact or elongated). These features collectively describe the spatial pattern of the data. For attribute statistical features, the system performs statistical analysis on all relevant attribute values ​​(which may be numerical or coded categorical), extracting indicators such as the mean (reflecting central tendency), variance (reflecting dispersion), and coefficient of variation (the ratio of variance to mean, used to compare the degree of variation of attributes at different scales). Each time an evaluation is required (before and after the update), the system will traverse all relevant entities, calculate the aforementioned spatial and attribute features of each entity, and then average the feature values ​​of all entities (or use other aggregation statistical methods) to finally obtain a multi-dimensional mean feature vector that can represent the overall state of the dataset at that moment.

[0065] The feature vectors of the state before the update are obtained respectively (denoted as vector μ). P ) and the feature vector of the updated state (denoted as vector μ) QAfterwards, the system calculates the Euclidean distance between the two mean vectors. Euclidean distance is a classic method for measuring the straight-line distance between two points in a multidimensional space; here, it is used to quantify the distance the overall feature snapshot of the dataset moves in a high-dimensional space composed of spatial and attribute features. This distance value intuitively reflects the comprehensive shift in the overall features of the dataset (including average shape, average size, average attribute fluctuations, etc.) caused by the linked updates. Compared to relative entropy, which only focuses on the distribution of a single attribute, this Euclidean distance captures information about the joint changes of multiple features. The relative entropy and Euclidean distance are weighted and fused to form the final comprehensive consistency evaluation index. Relative entropy (D) and feature vector Euclidean distance (Dist) reflect consistency from different perspectives: relative entropy focuses on the differences in the distribution of key business attributes, while Euclidean distance focuses on the shift of multi-dimensional overall features; the two are complementary. Therefore, the system assigns a weight to each of these two indices (e.g., weights ω1 and ω2, and ω1+ω2=1). The weight allocation can be adjusted according to business priorities. For example, if more attention is paid to the stability of attribute distribution, ω1 can be increased; if more attention is paid to the overall changes in spatial patterns, ω2 can be increased. Then, the final fusion index value is calculated using a linear combination formula (e.g., comprehensive index = ω1 × D + ω2 × Dist). This comprehensive index value integrates both distribution differences and feature shifts, providing a more robust and comprehensive evaluation result than a single relative entropy. The system can set a new, more stringent judgment threshold for this comprehensive index; only when the comprehensive index value falls below this threshold is the update considered to have passed the highest level of consistency verification. This fusion method reduces reliance on a single evaluation method and improves the reliability and anti-interference capability of judging true consistency in complex scenarios.

[0066] By constructing a multi-dimensional feature vector that integrates spatial morphology and attribute statistics, and calculating the Euclidean distance between the vectors before and after the update, the descriptive dimensions for the overall state changes of the dataset are enriched. Furthermore, by weighting and fusing relative entropy with this distance, a more comprehensive, robust, and configurable integrated consistency evaluation index is created. This index can more sensitively and accurately capture various potential impacts that linked updates may bring, effectively making up for the shortcomings of single statistical testing methods, and significantly improving the system's ability and reliability to conduct multi-angle, in-depth, and automated evaluation of data update quality.

[0067] It should be noted that although the steps are described in a specific order above, this does not mean that they must be performed in that order. In fact, some of these steps can be executed concurrently, or even in a different order, as long as the required functionality is achieved. The number of devices and processing scale described herein are for simplification of the invention; applications, modifications, and variations of this invention will be readily apparent to those skilled in the art.

[0068] Although embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. They can be applied to various fields suitable for the present invention. For those skilled in the art, other modifications can be easily made. Therefore, without departing from the general concept defined by the claims and their equivalents, the present invention is not limited to the specific details and illustrations shown and described herein.

Claims

1. A knowledge-driven spatiotemporal data collaborative update method, characterized in that, Includes the following steps: Step 1: Analyze the natural resource business process and extract business rules. Use OWL2 language to convert the business rules into machine-executable rules. Calculate the spatial adjacency of geographic entities and the temporal correlation of geographic entity states based on the standard defined topological relationship. With geographic entities as core nodes, machine-executable rules, adjacency, and temporal correlation as edge attributes, construct a business-space-temporal knowledge graph. Step 2: Based on log analysis, time series comparison and spatial overlay technology, extract the attribute change increment of structured business data and the spatial change increment of spatial data, and calculate the comprehensive change degree of geographic entities based on the attribute change increment and spatial change increment. When the comprehensive change degree exceeds the preset threshold, it is determined that the geographic entity has undergone significant change. Step 3: When a geographic entity undergoes significant changes, a breadth-first search is performed based on the business-spatial-temporal knowledge graph to identify the set of associated geographic entities related to the geographic entity, and an event-driven model is used to trigger the synchronous update of the set of associated geographic entities. Step 4: Calculate the spatial overlap between related geographic entities during the synchronous update process. When the spatial overlap exceeds the preset overlap threshold, it is determined to be a spatial conflict. The spatial conflict is resolved based on business priority, and the consistency of data distribution before and after the update is evaluated using relative entropy. Step 5: Calculate the time delay for each level of regional nodes to complete the synchronization update, and generate a version number containing metadata about the update initiation time, geographic entity status, and operator to support version rollback.

2. The knowledge-driven spatiotemporal data collaborative update method according to claim 1, characterized in that, In step 1, the business rules are converted into machine-executable rules using the OWL2 language, and the quantization function used for the rule triggering conditions is defined as follows: ; in, a k ( ev ) is a business event ev The k One attribute, V k This is the set of trigger thresholds corresponding to this attribute. m The total number of attributes associated with the rule; quantified. This is a Boolean value; when its value is 1, it indicates that the rule has been triggered.

3. The knowledge-driven spatiotemporal data collaborative update method according to claim 2, characterized in that, In step 1, the specific method for calculating the spatial adjacency of geographic entities is as follows: Based on the standard-defined topological relationships, calculate the spatial entities in two geographic entities. S 1 and S The adjacency degree of 2 is calculated using the following formula: in, Representing spatial entities S The boundary, Indicates the boundary length; when adjacent degree When they are, they are determined to be adjacent geographical entities; The specific method for calculating the temporal correlation degree of geographic entity status is as follows: Calculate geographic entities e At the point of time t 1 and t 2. State correlation The calculation formula is: in, λ This is the time decay coefficient, with a value ranging from 0.01 to 0.1; When constructing a business-spatial-temporal knowledge graph, geographic entities are used as core nodes, and machine-executable rules are incorporated. Adjacency and time correlation As an edge attribute connecting geographic entity nodes, it is used for fusion modeling to obtain a unified representation of entity-relationship-rule.

4. The knowledge-driven spatiotemporal data collaborative update method according to claim 1, characterized in that, In step 2, the specific method for extracting the attribute change increments of structured business data is as follows: For structured business data, compare the hash values ​​of new and old fields to detect attribute changes. At that time, determine the attribute increment. ,in For hash functions, a new For new field data, a old This is old field data; The method for extracting spatial change increments from spatial data is as follows: Calculate two periods of spatial data S new and S old The symmetric difference between them yields the spatial variation increment. ΔS ,Right now: ,in Added data Reduced portion of the data; Calculate the overall degree of change of geographic entities C ( e The specific formula is as follows: in, α This represents the spatial weight, with a value ranging from 0.6 to 0.

8. Represents the area of ​​a space; indicator function In attribute increment The value is 1 if it is true, and 0 otherwise. w k For the first k Preset weights for each attribute, n Total number of attributes; The preset threshold for the overall degree of change is θ ,when At that time, determine the geographic entity e Significant changes have occurred.

5. The knowledge-driven spatiotemporal data collaborative update method according to claim 1, characterized in that, In step 3, a breadth-first search is performed based on the business-spatial-temporal knowledge graph to identify the set of associated geographic entities that are related to the geographic entity. E rel The specific identification and judgment conditions are as follows: in, ev To act on geographic entities e And business events that triggered significant changes in it, e' Indicates candidate associated geographic entities, conditions , , These correspond to spatial adjacency, temporal association, and business rule association between two geographical entities, respectively. Synchronous updates are triggered using an event-driven model, specifically including: For the set of associated geographic entities E rel Each geographic entity in e' Calculate update priority P ( e' The calculation formula is: in, β and γ The weighting coefficients and β+γ≤1 , The association weights are assigned based on business rules; Based on the update priority of each geographic entity, the synchronous update of the associated geographic entities is triggered and executed.

6. The knowledge-driven spatiotemporal data collaborative update method according to claim 1, characterized in that, In step 4, the geographic entities are calculated using the following formula. e 1 and e 2 Between: When spatial overlap At that time, it was determined to be a spatial conflict; The process of resolving spatial conflicts based on business priorities specifically includes: Calculate business priorities for conflicting geographic entities P bus ( e The calculation formula is: in, b k Representing geographic entities e The associated first k Item business type, level ( b k This indicates the preset level for this service type. w k The preset weight for this business type, m This represents the total number of associated business types. When conflicts are resolved, the update results of geographic entities with higher business priority will override the update results of geographic entities with lower business priority.

7. The knowledge-driven spatiotemporal data collaborative update method according to claim 6, characterized in that, In step 4, relative entropy is used to evaluate the consistency of the data distribution before and after the update, specifically including: Calculate the data distribution before the update P Compared with the updated data distribution Q The relative entropy between : in, N The total number of values ​​that the data can take; when If the value is less than the preset consistency threshold, the update is deemed to meet the consistency requirements.

8. The knowledge-driven spatiotemporal data collaborative update method according to claim 1, characterized in that, In step 5, the time delay for each level of node to complete the synchronization update is calculated. T delay The specific calculation formula is as follows: in, t 0 represents the update initiation time. t 1, t 2, t 3, ... represent the time when the provincial, municipal, and county-level nodes completed the synchronization update, respectively; Generate a version number that includes the update initiation time, geographic entity status, and operator metadata. V The specific generation method is as follows: in, e For identifying geographical entities that have undergone significant changes, E rel For the identification of a set of associated geographic entities, It is the hyperbolic tangent function.

9. The knowledge-driven spatiotemporal data collaborative update method according to claim 1, characterized in that, In step 3, the event-driven model used specifically includes event listeners, event resolvers, and update schedulers, where: The event listener monitors for significant changes to geographic entities in real time and encapsulates the events into standard event objects containing geographic entity identifiers, change types, and timestamps. The event parser receives standard event objects, extracts geographic entity identifiers from them, and performs a breadth-first search based on the business-spatial-temporal knowledge graph to identify the set of associated geographic entities. At the same time, it parses the change type to determine the type of update operation to be triggered. The update scheduler calculates the update priority of each associated entity based on the set of associated geographic entities output by the event resolver and the update operation type, combined with the edge attributes in the knowledge graph. It then generates an update task queue according to the priority order and schedules and executes the update tasks in the queue in an asynchronous and non-blocking manner.

10. A knowledge-driven spatiotemporal data collaborative update method according to claim 7, characterized in that, In the process of using relative entropy to assess the consistency of data distribution before and after the update, a comprehensive consistency assessment index is also generated, which specifically includes: In calculating relative entropy Simultaneously, spatial distribution features and attribute statistical features are extracted based on the spatial geometric shape and attribute value set of geographic entities, respectively. Spatial distribution features include the logarithm of entity area, the logarithm of entity perimeter, and entity shape index; attribute statistical features include the mean, variance, and coefficient of variation of all attribute values. The extracted spatial distribution features and attribute statistical features are combined in a preset order to form a multi-dimensional feature vector representing the state of geographic entity data. The mean vectors of all geographic entity corresponding feature vectors in the data distribution P before the update and the data distribution Q after the update are calculated respectively. The Euclidean distance between the two mean vectors is calculated, and this Euclidean distance is compared with the relative entropy. We perform weighted fusion to obtain a comprehensive consistency evaluation index.

Images