Community governance decision generation method and device based on knowledge graph

Abstract

The embodiment of the invention discloses a community governance decision generation method and device based on a knowledge graph. The method comprises the following steps: collecting community governance data; performing information extraction and text processing on the community governance data by adopting a data extraction model to generate structured data; constructing a causal analysis atlas, a decision treatment atlas and a space-time correlation atlas by using the structured data; fusing the causal analysis graph, the decision governance graph and the space-time association graph to obtain a community governance knowledge graph, wherein the community governance knowledge graph supports semantic-based query and reasoning; acquiring resident appeal information, and performing risk reasoning based on a preset reasoning rule to obtain a risk reasoning result; and based on a risk reasoning result, in combination with spatio-temporal information defined by a spatio-temporal association graph, querying in the community governance knowledge graph, and generating a community governance scheme according to a query result. According to the invention, the dynamism, the accuracy and the intelligent level of community management can be improved.

Description

Technical Field

[0001] This application relates to the fields of knowledge graph technology and intelligent decision support, and in particular, to a method and device for generating community governance decisions based on a knowledge graph. Background Art

[0002] With the acceleration of the urbanization process, community governance faces challenges such as diverse resident demands, complex data sources, and dynamic governance needs. Traditional community governance decision support methods mostly rely on static data analysis and have the following problems: Lack of dynamics: Existing methods are difficult to process multi-dimensional data such as resident demands, time information, and geographical information in real time, resulting in lagged decisions; Difficult knowledge integration: Information such as resident demands, government responsibilities, and spatio-temporal data is scattered, lacking a unified modeling and analysis framework; Limited causal analysis ability: Existing methods cannot effectively mine the causal relationships among demands, risks, and governance measures, and the decisions lack accuracy and scientificity.

[0003] In the prior art, although some governance methods based on big data can provide decision support through statistical analysis, they lack in-depth modeling of spatio-temporal dimensions and causal relationships and are difficult to meet the complex needs of modern community governance. Therefore, there is an urgent need for a method that can integrate multi-source data, support risk reasoning, and dynamic decision-making. Summary of the Invention

[0004] This application provides a method and device for generating community governance decisions based on a knowledge graph, achieving efficient decision generation.

[0005] This application provides the following solutions:

[0006] According to a first aspect, a method for generating community governance decisions based on a knowledge graph is provided. The method includes: collecting community governance data, where the community governance data includes residents' demands, geographical information, time information, and government department responsibility information; using a data extraction model to perform information extraction and text processing on the community governance data to generate structured data; using the structured data to construct a causal analysis graph, a decision-making governance graph, and a spatio-temporal association graph; the causal analysis graph is used to describe the relationships among demands, potential hazards, risks, events, and consequences; the decision-making governance graph is used to describe the relationships among demands, governance participants, governance solutions, and resource allocation; the spatio-temporal association graph is used to describe the relationships between demands and time data, and between demands and spatial data; fusing the causal analysis graph, the decision-making governance graph, and the spatio-temporal association graph to obtain a community governance knowledge graph, where the community governance knowledge graph supports semantic-based queries and inferences; obtaining residents' demand information, performing risk inference based on preset inference rules to generate a risk inference result; based on the risk inference result, combining with the spatio-temporal information defined by the spatio-temporal association graph, querying in the community governance knowledge graph, and generating a community governance solution according to the query result.

[0007] According to an implementable manner in an embodiment of the present application, the collecting of the community governance data further includes one or more of the following: realizing multi-source data integration through an API interface; cleaning, standardizing the format, and removing redundancy from the community governance data; using an automated script to perform data format conversion on the community governance data, and using an anomaly detection algorithm to remove low-confidence data.

[0008] According to an implementable manner in an embodiment of the present application, the construction process of the spatio-temporal association graph includes: using the OWL-Time temporal ontology to define the time attributes of the residents' demands, and constructing a time dimension graph according to the time attributes, where the time dimension graph includes the association relationship between the residents' demands and time information; using the GeoNames geographical ontology to define spatial entities and their relationships, and constructing a spatial dimension graph according to the spatial entities and their relationships, where the spatial dimension graph includes the association relationship between the residents' demands and location information.

[0009] According to an implementable manner in an embodiment of the present application, the data extraction model includes: a fine-tuned UIE model and a fine-tuned BERT model; the using of the data extraction model to perform information extraction and text processing on the community governance data includes: using the fine-tuned UIE model to perform information extraction on the community governance data, and using the fine-tuned BERT model to perform text processing on the community governance data.

[0010] According to an implementable manner in an embodiment of the present application, the fusion of the causal analysis graph, the decision-making governance graph, and the spatio-temporal association graph to generate a community governance knowledge graph includes: extracting key entities and their relationships in the community governance scenario according to the causal analysis graph, the decision-making governance graph, and the spatio-temporal association graph, and constructing an initial structure of the community governance knowledge graph; performing node embedding on the community governance knowledge graph based on a graph neural network, and using a graph convolutional network to update node features to generate entity feature representations; adopting an attention mechanism to enhance the feature expression ability of associated entities, and performing weighted update according to node similarity and semantic association; fine-tuning the graph neural network in combination with the community governance prediction task, and outputting the dynamic association pattern between complex entities in the community governance knowledge graph.

[0011] According to an implementable manner in an embodiment of the present application, the fusion of the causal analysis graph, the decision-making governance graph, and the spatio-temporal association graph to obtain a community governance knowledge graph includes: obtaining the causal relationship data defined in the causal analysis graph, including the semantic associations between resident demands, hidden dangers, risks, events, and consequences; obtaining the governance relationship data defined in the decision-making governance graph, including the collaboration rules between governance participants, governance plans, and resource allocation; obtaining the spatio-temporal relationship data defined in the spatio-temporal association graph, including the temporal attributes and spatial attributes of community events based on the time ontology and the geographical ontology; performing semantic alignment on the causal relationship data, the governance relationship data, and the spatio-temporal relationship data, and establishing cross-graph association rules, where the association rules include the mapping of causal relationships and spatio-temporal attributes, the connection of governance relationships and causal relationships, and the link of spatio-temporal attributes and governance relationships; based on the association rules after semantic alignment, fusing and generating the community governance knowledge graph, and the community governance knowledge graph is represented in the form of entity nodes and relationship edges.

[0012] According to an implementable manner in the embodiments of the present application, querying in the knowledge graph based on the risk reasoning result and combining the spatio-temporal information defined by the spatio-temporal association graph, and generating a community governance plan according to the query result includes: determining, according to the risk reasoning result, risk types and potential consequences related to residents' demands; using the time attribute defined by the spatio-temporal association graph to extract time information associated with the risk types from the knowledge graph, including the time point or time interval when the event occurs; using the spatial attribute defined by the spatio-temporal association graph to extract geographical location information associated with the risk types from the knowledge graph, including the community, block or specific facility where the event occurs; constructing spatio-temporal query conditions based on the time information and geographical location information, performing semantic queries in the community governance knowledge graph, and obtaining events, trends or governance resources that match the risk types in the spatio-temporal dimension; according to the query result, combining the governance participants and resource allocation rules defined by the decision-making governance graph, generating a community governance plan for the residents' demands, and the community governance plan includes governance measures and participating entities.

[0013] According to a second aspect, there is provided a method for generating a community governance decision based on a knowledge graph. The device includes: a data collection unit configured to collect community governance data, where the community governance data includes residents' demands, geographical information, time information, and government department responsibility information; a structured data generation unit configured to perform information extraction and text processing on the community governance data using a data extraction model to generate structured data; a sub-graph construction unit configured to use the structured data to construct a causal analysis graph, a decision-making governance graph, and a spatio-temporal association graph; the causal analysis graph is used to describe the relationship between demands, potential hazards, risks, events, and consequences; the decision-making governance graph is used to describe the relationship between demands, governance participants, governance plans, and resource allocation; the spatio-temporal association graph is used to describe the relationship between demands and time data, and between demands and spatial data; a graph fusion unit configured to fuse the causal analysis graph, the decision-making governance graph, and the spatio-temporal association graph to obtain a community governance knowledge graph, and the community governance knowledge graph supports semantic-based queries and inferences; a risk reasoning unit configured to obtain residents' demand information and perform risk reasoning based on preset reasoning rules to generate a risk reasoning result; a plan generation unit configured to query in the community governance knowledge graph based on the risk reasoning result and combine the spatio-temporal information defined by the spatio-temporal association graph, and generate a community governance plan according to the query result.

[0014] According to a third aspect, there is provided a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps of the method according to any one of the first aspects are implemented.

[0015] According to a fourth aspect, there is provided an electronic device, including:

[0016] One or more processors; and a memory associated with the one or more processors, the memory being used to store program instructions, and when the program instructions are read and executed by the one or more processors, the steps of the method according to any one of the above first aspects are executed.

[0017] According to the specific embodiments provided in the present application, the following technical effects are disclosed in the present application:

[0018] The present invention provides a method for generating community governance decisions based on a knowledge graph. By constructing a causal analysis graph, a decision governance graph, and a spatio-temporal association graph, and fusing them to generate a community governance knowledge graph, risk reasoning of residents' demands and generation of governance solutions are realized. The present invention can improve the dynamic, precise, and intelligent levels of community governance.

[0019] Of course, it is not necessary for any product implementing the present application to achieve all the above-mentioned advantages simultaneously. Description of the Drawings

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required to be used in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0021] Figure 1 It is a system architecture diagram applicable to the embodiments of the present application;

[0022] Figure 2 It is a flowchart of the method for generating community governance decisions based on a knowledge graph provided by the embodiments of the present application;

[0023] Figure 3 It is the fine-tuning parameters of the UIE model in the present application;

[0024] Figure 4 It is the fine-tuning parameters of the BERT model in the present application;

[0025] Figure 5 It is a schematic diagram of the causal analysis graph of the present application;

[0026] Figure 6 It is a schematic diagram of the decision governance graph of the present application;

[0027] Fig. 7(a) is a schematic diagram of the time dimension of the spatio-temporal association graph of the present application;

[0028] Fig. 7(b) is a schematic diagram of the space dimension of the spatio-temporal association graph of the present application;

[0029] Figure 8 This is a structural block diagram of a community governance decision generation device based on a knowledge graph provided by an embodiment of the present application;

[0030] Figure 9 This is a schematic block diagram of an electronic device provided by an embodiment of the present application. Detailed implementation manners

[0031] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art belong to the scope of protection of the present application.

[0032] The terms used in the embodiments of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The singular forms "a", "said", and "the" used in the embodiments of the present invention and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise.

[0033] It should be understood that the term " / and / " used herein is only a description of the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " herein generally represents an "or" relationship between the associated objects before and after.

[0034] Depending on the context, the word "if" as used herein may be interpreted as "when", "when...", "in response to determining", or "in response to detecting". Similarly, depending on the context, the phrase "if determined" or "if detecting (stated condition or event)" may be interpreted as "when determined", "in response to determining", "when detecting (stated condition or event)", or "in response to detecting (stated condition or event)".

[0035] The present application provides a method for generating community governance decisions based on a knowledge graph. For the convenience of understanding the present application, the system architecture on which the present application is based will be described first. Figure 1 An exemplary system architecture to which the embodiments of the present application can be applied is shown. As Figure 1 shown in, the system architecture may include a user device and a community governance decision generation device based on a knowledge graph located on the server side.

[0036] Among them, the user equipment may include but is not limited to, such as: intelligent mobile terminals, smart home devices, wearable devices, PCs (Personal Computers), etc. Among them, intelligent mobile devices may include, such as, mobile phones, tablet computers, laptop computers, PDAs (Personal Digital Assistants), Internet cars, etc. Smart home devices may include smart TVs, smart refrigerators, and so on. Wearable devices may include, such as, smart watches, smart glasses, virtual reality devices, augmented reality devices, mixed reality devices (i.e., devices that can support virtual reality and augmented reality), and so on.

[0037] The community governance decision generation device based on the knowledge graph can adopt the method provided in the embodiments of this application to generate decision results according to user demands. The community governance decision generation device based on the knowledge graph can be set as an independent server, or can be set in a server group, or can also be set in a cloud server. A cloud server, also known as a cloud computing server or a cloud host, is a host product in the cloud computing service system, which is used to solve the defects of difficult management and weak service scalability existing in traditional physical hosts and virtual private server (VPS, Virtual Private Server) services. In addition to Figure 1 the architecture shown, the community governance decision generation device based on the knowledge graph can also be set in a computer terminal with strong computing power.

[0038] Figure 2 The figure is a flowchart of a community governance decision generation method provided by the embodiments of this application. This method can be executed by the Figure 1 community governance decision generation device based on the knowledge graph in the system shown. As Figure 2 shown in, this method may include the following steps:

[0039] Step 201: Collect community governance data, where the community governance data includes residents' demands, geographical information, time information, and government department responsibility information.

[0040] Step 202: Use a data extraction model to perform information extraction and text processing on the community governance data to generate structured data.

[0041] Step 203: Use the structured data to construct a causal analysis graph, a decision governance graph, and a spatio-temporal association graph; the causal analysis graph is used to describe the relationships among demands, potential hazards, risks, events, and consequences; the decision governance graph is used to describe the relationships among demands, governance participants, governance solutions, and resource allocation; the spatio-temporal association graph is used to describe the relationships between demands and time data, and between demands and spatial data.

[0042] Step 204: Integrate the causal analysis graph, the decision-making governance graph, and the spatio-temporal association graph to obtain a community governance knowledge graph, which supports semantic-based querying and reasoning.

[0043] Step 205: Obtain residents' demand information, perform risk reasoning based on preset reasoning rules, and obtain a risk reasoning result.

[0044] Step 206: Based on the risk reasoning result, combined with the spatio-temporal information defined by the spatio-temporal association graph, query in the community governance knowledge graph, and generate a community governance plan according to the query result.

[0045] As can be seen from the above process, the present invention provides a method for generating community governance decisions based on a knowledge graph. By constructing a causal analysis graph, a decision-making governance graph, and a spatio-temporal association graph, and integrating them to generate a community governance knowledge graph, it realizes risk reasoning of residents' demands and generation of governance plans. The present invention can improve the dynamic, accurate, and intelligent levels of community governance.

[0046] First, the above step 201, that is, "collect community governance data, where the community governance data includes residents' demands, geographical information, time information, and government department responsibility information", will be described in detail in combination with an embodiment.

[0047] This step obtains the original data required for community governance from multiple sources, providing materials for subsequent structured processing and knowledge graph construction. Community governance data is a multi-dimensional information set reflecting the operation status and governance needs of the community. Its collection process needs to cover different data types and sources to ensure the comprehensiveness and representativeness of the data. Specifically, residents' demands refer to the needs or problems raised by community residents, usually presented in text form, such as "there is an illegal construction problem in a certain community"; geographical information includes spatial location data within the community, such as the longitude and latitude coordinates of a community, block, or specific facility, and can also include administrative division data; time information records the time point or interval when an event or demand occurs, such as "at 2025-01-05 morning"; government department responsibility information covers the functions and responsibility divisions of relevant departments, such as "the comprehensive law enforcement team is responsible for illegal construction governance".

[0048] The collection of these data is achieved through various means, including but not limited to API interfaces, database queries, and manual entry, aiming to provide comprehensive data support for community governance. In actual implementation, appropriate methods and technologies need to be adopted for different data types during the data collection process. Taking residents' demands as an example, they can be collected through the online submission system or phone hotline of the community management platform. For instance, a resident submits a demand of "The illegal construction at the entrance of Community X affects traffic" through the APP, and the system records it as text data. The collection of geographical information can utilize GIS (Geographic Information System) tools or public geographical databases (such as GeoNames) to obtain the specific location of Community X, such as latitude and longitude coordinates (latitude 39.9042, longitude 116.4074) and its administrative affiliation (such as "M Sub-district"). The collection of time information is usually bound to demands or events and is directly extracted from the submission timestamp, such as "2025-01-05T08:00:00+08:00", or the time range of the event occurrence is obtained through log records. The collection of government department responsibility information relies on official databases or documents. For example, the information of "The comprehensive law enforcement team is responsible for illegal construction removal and punishment" is extracted from the responsibility manual of the urban management department. Through API interfaces, cross-departmental data integration can be achieved. For example, the traffic impact data can be obtained by calling the API of the traffic department, and the responsibility assignment can be obtained by calling the API of the law enforcement department.

[0049] After obtaining the community governance data, it is also possible to clean, standardize the format, and remove redundancy from the community governance data. Community governance data may contain noise, non-standard formats, or duplicate information during the collection process. For example, spelling mistakes in residents' demands, inconsistent expressions in geographical information, or multiple recording methods of time information. Data cleaning improves the accuracy of data by removing this noise data (such as the misspelling of "illegal construction" as "illegal health"); format standardization unifies heterogeneous data into a standard representation. For example, "5th January in the morning" is standardized to "2025-01-05T08:00:00+08:00"; redundancy removal deletes duplicate records, such as the situation where the same demand is submitted multiple times. These operations ensure the cleanliness and consistency of the data.

[0050] Automated scripts can also be used to convert the data format of community governance data, and anomaly detection algorithms can be used to remove low-confidence data. Automated scripts are used to convert the cleaned data into a format suitable for subsequent processing. For example, text data is converted into JSON or RDF structures, while anomaly detection algorithms can identify and eliminate low-confidence data based on statistical or machine learning methods (such as the Isolation Forest algorithm). For example, geographical coordinates or timestamps with a confidence score lower than 0.5 are removed, thereby further optimizing the data quality.

[0051] The above step 202, that is, "using a data extraction model to perform information extraction and text processing on the community governance data to generate structured data", will be described in detail below in combination with the embodiments.

[0052] This step transforms the original community governance data into a structured form that can be used for subsequent graph construction and reasoning. Community governance data usually includes residents' demands, geographical information, time information, and government department responsibility information, and these data often exist in unstructured or semi-structured forms, such as natural language texts (such as "residents of a certain community reported the illegal construction problem"), tables, or log records. The data extraction model resolves these heterogeneous data into a unified structured representation through two major functions: information extraction and text processing, so as to support the construction of the knowledge graph. Specifically, information extraction is responsible for identifying key entities and relationships in the data. For example, entities "illegal construction" and relationship "causing risks" are extracted from the residents' demand text; while text processing performs semantic analysis and normalization on the data. For example, "illegal construction in a certain community on January 5th" is processed into standard time and location annotations. This process relies on advanced natural language processing (NLP) technologies, and usually uses a fine-tuned deep learning model to improve the accuracy of extraction and processing.

[0053] As an implementable method, the data extraction model can include two sub-modules: one for information extraction and the other for text processing. The fine-tuned UIE (Unified Information Extraction) model is used for information extraction. The UIE model can extract entities, relationships, and events from text. For example, from the demand "residents of a certain community reported the illegal construction problem and reported it on January 5th", triples <demand 1, hasEntity, illegal construction> and <demand 1, hasTime, January 5th> are extracted, and its accuracy can reach above 0.79 through fine-tuning. Among them, the fine-tuning parameters of the UIE model are as Figure 3 shown. The traditional UIE model performs extraction directly without training, while in this application, different training rounds are carried out for UIE. When the number of training rounds reaches the 20th round, the effect of the fine-tuned model has reached the best.

[0054] At the same time, the fine-tuned BERT (Bidirectional Encoder Representations from Transformers) model is used for text processing. Through context semantic analysis, BERT transforms non-standard expressions (such as "the illegal construction in the community is very serious") into standardized structured data, such as <demand 1, hasDescription, serious illegal construction>, and the accuracy usually exceeds 0.8. Among them, the fine-tuning parameters of the BERT model are as Figure 4As shown in the figure, BERT has been trained many times, with the highest accuracy at round 20. The fine-tuning process is carried out according to the specific corpus in the field of community governance. For example, the model parameters are adjusted using labeled appeal data (such as the learning rate is set to 1e-5 and the batch size is 16) to adapt to the language characteristics of the governance scenario.

[0055] For example, suppose the input community governance data is a text of residents' demands: "Residents of X community reported that there were illegal buildings at the entrance of the community on the morning of January 5, 2025, which affected traffic." First, the UIE model extracts information from the text, identifies the entities "X community", "illegal buildings", "the morning of January 5, 2025", and the relationship "affecting traffic", and generates preliminary structured data: <Appeal 1, hasLocation, X community>, <Appeal 1, hasEntity, illegal buildings>, <Appeal 1, hasTime, 2025-01-05T08:00:00>, <Appeal 1, hasImpact, traffic impact>. Subsequently, the BERT model processes the text, analyzes the semantics of "affecting traffic", normalizes it into a standard description, and supplements potential attributes, such as <Appeal 1, hasSeverity, high> (based on the severity of "affecting traffic"). The final output structured data is a set of standardized triples, which can be directly used for subsequent causal analysis maps and spatiotemporal association map construction.

[0056] The following is a detailed description of step 203 in conjunction with an embodiment, namely, "using the structured data to construct a cause-effect analysis map, a decision governance map and a spatiotemporal association map; the cause-effect analysis map is used to describe the relationship between demands, hidden dangers, risks, events and consequences; the decision governance map is used to describe the relationship between demands, governance participants, governance plans and resource allocation; the spatiotemporal association map is used to describe the relationship between demands and time data, and between demands and spatial data."

[0057] This step converts structured data into three types of sub-graphs, providing a basis for the subsequent integration and reasoning of community governance knowledge graphs. Structured data is usually stored in key-value pairs, triples, or other standard formats, and contains information such as residents' demands, geographic information, time information, and government department responsibilities. Through this step, the data is organized into a graph structure with specific semantic relationships, focusing on causal logic, governance collaboration, and spatiotemporal characteristics. Each sub-graph represents knowledge of a specific dimension in the form of entities (nodes) and relationships (edges), supporting multi-dimensional analysis of community governance.

[0058] Specifically, the construction of the causal analysis graph uses the demands and their related information in structured data to describe the causal chain among demands, potential hazards, risks, events, and consequences. For example, from the structured data <Demand 1, hasEntity, illegal construction> and <Demand 1, hasImpact, traffic impact>, the entity "illegal construction" is extracted as a potential hazard, "violation of regulations" as a risk, and "traffic impact" as a consequence, and the graph relationships <illegal construction, causes, violation of regulations> and <violation of regulations, leadsTo, traffic impact> are constructed.

[0059] The decision-making and governance graph focuses on the governance process and describes the relationships among demands, governance participants, governance solutions, and resources based on the responsibility and demand information in the data. For example, from <Demand 1, hasEntity, illegal construction> and <Law enforcement team, hasDuty, illegal construction governance>, <Demand 1, addressedBy, law enforcement team>, <Law enforcement team, executes, investigation and punishment>, and <Investigation and punishment, requires, law enforcement resources> are constructed.

[0060] The spatio-temporal association graph uses time and space data to describe the associations between demands and time and space. For example, from <Demand 1, hasTime, 2025-01-05T08:00:00> and <Demand 1, hasLocation, Community X>, <Demand 1, occursAt, 2025-01-05T08:00:00> and <Demand 1, locatedIn, Community X> are constructed.

[0061] As an implementable method, the construction process of the spatio-temporal association graph includes: using the OWL-Time time ontology to define the time attributes of the resident demands, constructing a time dimension graph according to the time attributes, and the time dimension graph includes the association relationship between the resident demands and time information; using the GeoNames geographical ontology to define the spatial entities and their relationships, and constructing a spatial dimension graph according to the spatial entities and their relationships, and the spatial dimension graph includes the association relationship between the resident demands and location information.

[0062] Specifically, in the construction process, the time attributes of residents' demands are first defined using the OWL-Time temporal ontology. OWL-Time is a standardized time representation framework that supports semantic descriptions of time points (such as "08:00 on January 5, 2025"), time intervals (such as "morning rush hour"), and periods (such as "winter"). Time information is extracted from structured data. For example, the demand "illegal construction in Community X on January 5" is parsed as <Demand 1, hasTime, 2025-01-05T08:00:00> and represented based on OWL-Time as <Demand 1, time:occursAt, "2025-01-05T08:00:00"^^xsd:dateTime>. According to these time attributes, a temporal dimension graph is constructed, which includes nodes (such as "Demand 1", "2025-01-05T08:00:00") and edges (such as "occursAt"), forming an association relationship network between demands and time information. The temporal dimension graph not only records a single time point but can also be extended to periodic features, such as <Demand 1, occursIn, winter>, to support the analysis of seasonal trends. Next, the GeoNames geographical ontology is used to define spatial entities and their relationships. GeoNames is a global geographical database that provides semantic representations of place names, coordinates, and administrative affiliations. For example, "Community X" can be mapped to <Community X, gn:name, "Community X">, <Community X, wgs84_pos:lat, 39.9042>, <Community X, wgs84_pos:long, 116.4074>. Location information is extracted from structured data, such as <Demand 1, hasLocation, Community X>, and relationships between spatial entities are constructed based on GeoNames. For example, <Community X, governedBy, Street M>. According to these spatial entities and their relationships, a spatial dimension graph is constructed, which includes nodes (such as "Demand 1", "Community X") and edges (such as "locatedIn", "governedBy"), forming an association relationship network between demands and location information. The spatial dimension graph not only represents specific locations but can also include hierarchical relationships, such as the subordination of a block to an administrative region.

[0063] For example, assume that the structured data is: <Appeal 1, hasEntity, illegal construction>, <Appeal 1, hasTime, 2025-01-05T08:00:00>, <Appeal 1, hasLocation, Community X>, <Appeal 1, hasImpact, traffic impact>, <Law Enforcement Team, hasDuty, illegal construction governance>. The causal analysis graph construction process identifies "illegal construction" as a hidden danger, "regulation violation" as a risk, and "traffic impact" as a consequence, forming triples <illegal construction, causes, regulation violation> and <regulation violation, leadsTo, traffic impact>, which constitute a causal chain graph. The decision-making governance graph extracts governance-related entities from the same appeal and duty data, generating <Appeal 1, addressedBy, Law Enforcement Team> and <Law Enforcement Team, executes, investigation and punishment>, representing the governance process. The spatio-temporal association graph generates <Appeal 1, occursAt, 2025-01-05T08:00:00> and <Appeal 1, locatedIn, Community X> based on time and space attributes, forming a spatio-temporal relationship network. These sub-graphs are stored in RDF format, such as <ex:Appeal1, ex:causes, ex:Risk_RegulationBreach> (causality), <ex:Appeal1, ex:addressedBy, ex:EnforcementTeam> (governance), <ex:Appeal1, time:occursAt, "2025-01-05T08:00:00"^^xsd:dateTime> (spatio-temporal).

[0064] As an implementable approach, this application visually presents schematic diagrams of three types of sub-graphs, such as Figure 5 shown as the schematic diagram of the causal analysis graph,[[]] Figure 6 as the schematic diagram of the decision-making governance graph; Figure 7(a) is the schematic diagram of the time dimension in the spatio-temporal association graph; Figure 7(b) is the schematic diagram of the space dimension in the spatio-temporal association graph.

[0065] The following describes in detail the above step 203, that is, "fusing the causal analysis graph, the decision-making governance graph, and the spatio-temporal association graph to obtain a community governance knowledge graph, and the community governance knowledge graph supports semantic-based query and reasoning", in combination with embodiments.

[0066] This step merges three types of sub-graphs (causal analysis graph, decision governance graph, and spatio-temporal association graph) into a unified community governance knowledge graph to achieve collaborative analysis of multi-dimensional data and intelligent decision-making support. Each sub-graph focuses on causal relationships, governance processes, and spatio-temporal characteristics respectively. The fusion process generates a comprehensive graph structure by integrating the entities and relationships of these graphs, including nodes (entities, such as demands, risks, law enforcement teams) and edges (relationships, such as "causes", "executes"). The fused community governance knowledge graph not only retains the semantic information of each sub-graph, but also achieves richer semantic expressions through newly added cross-graph associations, supporting semantic-based queries (such as SPARQL queries) and inferences (such as risk prediction), thus providing comprehensive data support and decision-making basis for community governance.

[0067] In the specific implementation, multiple methods can be adopted for the fusion process. As an implementable approach, according to the causal analysis graph, the decision governance graph, and the spatio-temporal association graph, key entities and their relationships in the community governance scenario are extracted to construct the initial structure of the community governance knowledge graph; node embedding of the community governance knowledge graph is performed based on the Graph Neural Network (GNN), and the node features are updated using the Graph Convolutional Network (GCN) to generate entity feature representations; the attention mechanism is adopted to enhance the feature expression ability of associated entities, and weighted updates are performed according to node similarity and semantic association; the graph neural network is fine-tuned in combination with the community governance prediction task to output the dynamic association patterns between complex entities in the community governance knowledge graph.

[0068] The core of this method lies in mining the deep associations between sub-graphs through machine learning techniques to generate an optimized knowledge graph that supports semantic queries and reasoning. In the specific implementation, entities and relationships are first extracted from three sub-graphs. For example, the causal analysis graph provides <illegal construction, causes, violation of regulations>, the decision-making and governance graph provides <law enforcement team, executes, investigation and punishment>, and the spatio-temporal association graph provides <appeal 1, locatedIn, Community X>. An initial knowledge graph is constructed, which contains nodes (such as "illegal construction" and "law enforcement team") and edges (such as "causes" and "executes"). Subsequently, the GNN generates embedding vectors for each node through GCN. For example, "illegal construction" is encoded as a 128-dimensional vector, and its features are updated based on neighbor nodes (such as "violation of regulations") to capture the graph structure information. The attention mechanism further enhances the key associations. For example, higher weights are assigned to "illegal construction" and "law enforcement team" because they have strong semantic relevance in the governance scenario. In the fine-tuning stage, community governance tasks (such as predicting high-risk areas) are used to optimize the model, and dynamic patterns are output, such as "illegal construction in Community X in winter needs to be prioritized for governance". The fused knowledge graph is represented by feature vectors and optimized relationships, supporting complex queries and predictions.

[0069] For example, assume the input data is: causal graph <illegal construction, causes, violation of regulations>, governance graph <law enforcement team, executes, investigation and punishment>, spatio-temporal graph <appeal 1, occursAt, 2025-01-05T08:00:00>. The initial graph contains these triples. The GNN embeds "illegal construction" as a vector [0.1, 0.5,...], fuses the features of "violation of regulations" after updating through GCN, the attention mechanism highlights the "illegal construction - law enforcement team" association, and finally outputs the pattern "High risk of illegal construction in Community X in winter, requiring the law enforcement team to handle". This method is applicable to scenarios that require in-depth mining of potential relationships, such as predicting risk areas not explicitly recorded. Its advantage lies in automated feature learning, but it has low interpretability and depends on training data and model tuning.

[0070] As an implementable approach, obtain the causal relationship data defined in the causal analysis graph, including the semantic associations among residents' demands, potential hazards, risks, events, and consequences; obtain the governance relationship data defined in the decision-making and governance graph, including the collaboration rules among governance participants, governance solutions, and resource allocation; obtain the spatio-temporal relationship data defined in the spatio-temporal association graph, including the temporal attributes and spatial attributes of community events based on the time ontology and geographical ontology; semantically align the causal relationship data, the governance relationship data, and the spatio-temporal relationship data to establish cross-graph association rules, where the association rules include the mapping of causal relationships to spatio-temporal attributes, the connection of governance relationships to causal relationships, and the linking of spatio-temporal attributes to governance relationships; based on the semantically aligned association rules, fuse and generate the community governance knowledge graph, which is represented in the form of entity nodes and relationship edges.

[0071] In a specific implementation, extract structured data from sub-graphs. For example, the causal graph provides <Illegal construction, causes, Violation of regulations>, the governance graph provides <Law enforcement team, executes, Investigation and punishment>, and the spatio-temporal graph provides <Appeal 1, occursAt, 2025-01-05T08:00:00> and <Appeal 1, locatedIn, Community X>. The semantic alignment process establishes association rules: <Violation of regulations, occursAt, 2025-01-05T08:00:00> (causality and spatio-temporal), <Investigation and punishment, addresses, Violation of regulations> (governance and causality), <Law enforcement team, operatesIn, Community X> (spatio-temporal and governance). Based on these rules, fuse and generate a knowledge graph containing triples <Appeal 1, causes, Violation of regulations>, <Violation of regulations, occursAt, 2025-01-05T08:00:00>, <Law enforcement team, executes, Investigation and punishment>, and store it in the RDF format (such as <ex:Appeal1, ex:causes, ex:Risk_RegulationBreach>). The graph supports SPARQL queries, such as "SELECT?solution WHERE {?risk ex:occursAt '2025-01-05'; ex:addressedBy?solution}", and returns "Investigation and punishment".

[0072] In the community governance scenario, the association rules after semantic alignment are derived from the data integration of three sub-graphs. Suppose the causal analysis graph provides <illegal construction, causes, violation of regulations>, the decision-making governance graph provides <law enforcement team, executes, investigation and punishment>, and the spatio-temporal association graph provides <appeal 1, occursAt, 2025-01-05T08:00:00> and <appeal 1, locatedIn, X community>. The semantic alignment process analyzes the semantic meanings of these data and establishes cross-graph association rules. For example, "the mapping of causal relationship and spatio-temporal attributes" generates <violation of regulations, occursAt, 2025-01-05T08:00:00>, indicating the time background of the risk; "the connection of governance relationship and causal relationship" generates <investigation and punishment, addresses, violation of regulations>, indicating that the governance plan targets specific risks; "the link of spatio-temporal attributes and governance relationship" generates <law enforcement team, operatesIn, X community>, clarifying the spatial scope of the governance subject. These rules are not simple splicing, but logical associations generated through semantic analysis (such as entity matching, relationship reasoning), ensuring that the fused knowledge graph contains both sub-graph information and new cross-domain connections.

[0073] For example, the input data is: causal graph <illegal construction, causes, violation of regulations>, governance graph <law enforcement team, executes, investigation and punishment>, and spatio-temporal graph <appeal 1, locatedIn, X community>. After alignment, the rules <violation of regulations, locatedIn, X community> and <investigation and punishment, addresses, violation of regulations> are generated and fused into the graph <appeal 1, causes, violation of regulations> - <violation of regulations, locatedIn, X community> - <law enforcement team, executes, investigation and punishment>. Querying "the governance plan for the risks in X community" returns "the law enforcement team conducts investigation and punishment", and inferring "if there is illegal construction during peak hours, then it is urgent" yields a priority recommendation.

[0074] The following describes in detail step 205 above, that is, "obtain residents' appeal information, perform risk reasoning based on preset reasoning rules, and generate risk reasoning results" in combination with embodiments.

[0075] Residents' appeal information is usually newly input governance requirements, such as "there is an illegal construction problem in a certain community". By combining with the structured information in the knowledge graph and analyzing based on preset reasoning rules. Preset reasoning rules are a set of formal logical judgments, usually defined in the form of "if - then", such as "if there is illegal construction, then the risk is violation of regulations". These rules are derived from causal analysis graphs or domain expert knowledge. The risk reasoning process is executed in the knowledge graph. By matching the appeal information with the graph data, the risk type and possible consequences are deduced, and finally the risk reasoning result is generated to support precise decision-making in community governance.

[0076] In a specific implementation, first, information on residents' demands is obtained, which can be input in real time through the community management platform. For example, the text "Residents of Community X reported that the illegal construction at the community entrance affects traffic". This information is parsed into a structured form, such as <Demand 1, hasEntity, illegal construction> and <Demand 1, hasImpact, traffic impact>, and is then connected to the community governance knowledge graph. Preset inference rules may include: ① "If the demand contains illegal construction, then the risk = violation of regulations"; ② "If the risk affects traffic, then the consequence = safety hazard". The inference process uses the data in the graph and a rule engine (such as SPARQL query or SWRL rules) for matching and derivation. For example, querying the graph finds <illegal construction, causes, violation of regulations>, and Rule ① is triggered, resulting in "risk = violation of regulations"; further matching <Demand 1, hasImpact, traffic impact>, Rule ② is triggered, resulting in "consequence = safety hazard". The final risk inference result is "The risk of Demand 1 is violation of regulations, which may lead to safety hazards", and it is output in a structured form, such as <Demand 1, hasRisk, violation of regulations> and <Demand 1, hasConsequence, safety hazard>.

[0077] As an implementable method, the risk inference result in this application can also include a risk level. The risk level is a quantitative or qualitative description of the degree of risk impact or urgency, such as "high", "medium", or "low", aiming to provide a more granular decision-making basis for community governance. By introducing the risk level, the generation of governance plans can be more targeted. For example, high-risk events are given priority, thus optimizing resource allocation and response efficiency.

[0078] In a specific implementation, the risk reasoning process starts from obtaining residents' demand information. For example, "On the morning of January 5, 2025, illegal construction at the entrance of X Community affected traffic" is parsed into structured data <Demand 1, hasEntity, illegal construction>, <Demand 1, hasTime, 2025-01-05T08:00:00>, <Demand 1, hasImpact, traffic impact>. The preset reasoning rules are not limited to identifying risk types and consequences, but also include evaluation rules for risk levels. For example, the rules can be defined as: ① "If the demand contains illegal construction, then the risk = violation of regulations"; ② "If the risk affects traffic, then the consequence = safety hazard"; ③ "If the event occurs during the traffic peak period (such as 8:00 in the morning), then the risk level = high; if it occurs during the non-peak period, then the risk level = medium". The reasoning process is executed in the community governance knowledge graph. Matching <illegal construction, causes, violation of regulations> (Rule ①), the risk type is obtained; matching <Demand 1, hasImpact, traffic impact>, the consequence "safety hazard" is obtained (Rule ②); combining <Demand 1, occursAt, 2025-01-05T08:00:00>, since 8:00 is the peak period, Rule ③ is triggered, and "risk level = high" is obtained. The final risk reasoning result is <Demand 1, hasRisk, violation of regulations>, <Demand 1, hasConsequence, safety hazard>, <Demand 1, hasRiskLevel, high>.

[0079] The following combines the embodiments to describe in detail the above step 206, that is, "Based on the risk reasoning result, combined with the spatio-temporal information defined by the spatio-temporal association graph, query in the community governance knowledge graph, and generate a community governance plan according to the query result".

[0080] The risk reasoning result provides the potential risks and consequences of the demand. For example, "Violation of regulations may lead to safety hazards", and the time and space attributes (such as the time point and location where the event occurs) defined by the spatio-temporal association graph provide specific context for the risks. By performing semantic queries in the knowledge graph and combining the governance rules in the decision-making governance graph, the system can match risks with governance resources and generate a plan including governance measures and participating entities.

[0081] In specific implementation, this step first determines the risk type and consequences based on the risk reasoning results. For example, assume the reasoning results are <Appeal 1, hasRisk, violation of regulations> and <Appeal 1, hasConsequence, safety hazard>. Then, spatio-temporal information is extracted from the spatio-temporal association graph, such as <Appeal 1, occursAt, 2025-01-05T08:00:00> (time) and <Appeal 1, locatedIn, X Community> (location). In the community governance knowledge graph, these information are integrated using SPARQL query to retrieve governance solutions and participants related to the risk. For example, the query statement is: SELECT?solution?participant WHERE {?risk rdf:type ex:RegulationBreach; ex:occursAt "2025-01-05T08:00:00"; ex:locatedIn "X Community"; ex:addressedBy?solution.?participant ex:executes?solution}, and the returned results may be <Investigation and punishment, addresses, violation of regulations> and <Law enforcement team, executes, investigation and punishment>. According to the query results and combined with the rules of the decision-making governance graph (such as "violation of regulations requires investigation and punishment by the law enforcement team"), a governance solution is generated: the measure is "conduct investigation and punishment on the illegal construction in X Community", and the participating entity is "law enforcement team".

[0082] For another example, assume the resident's appeal is "serious garbage accumulation in Y Block on January 6, 2025", and the risk reasoning results are <Appeal 2, hasRisk, sanitation hazard> and <Appeal 2, hasConsequence, disease transmission>. <Appeal 2, occursAt, 2025-01-06T09:00:00> and <Appeal 2, locatedIn, Y Block> are extracted from the spatio-temporal association graph. Query in the knowledge graph: SELECT?solution?participant WHERE {?risk rdf:type ex:SanitationHazard; ex:occursAt "2025-01-06T09:00:00"; ex:locatedIn "Y Block"; ex:addressedBy?solution.?participant ex:executes?solution}, and <Clean-up, addresses, sanitation hazard> and <Sanitation team, executes, clean-up> are returned. According to the results, a solution is generated: "clean up the garbage accumulation in Y Block, to be executed by the sanitation team". If historical trends are found during the query (such as "high incidence of sanitation hazards in Y Block in winter"), the solution can be further optimized by adding a measure of "regular inspections".

[0083] For another example, the risk reasoning result shows that the resident's appeal "There is an illegal construction at the entrance of Ruyuan North District in X Town" is most likely to trigger the risk of "violating regulations". The spatio-temporal information clarifies that the event location is "the entrance of Ruyuan North District in X Town" and the administrative jurisdiction is "M Sub-district", while the governance methods and participant information have been embedded in the knowledge graph. In the specific implementation, first, based on the risk reasoning result, the risk type is determined as "violating regulations", which is obtained through the aforementioned steps of "obtaining resident appeal information and conducting risk reasoning based on preset reasoning rules". For example, the rule is "If the appeal contains illegal construction, then the risk = violating regulations". Combining with the definition of the spatio-temporal association graph, the spatio-temporal information is extracted: the location is "the entrance of Ruyuan North District in X Town" (which can be expressed as <Appeal 1, locatedIn, the entrance of Ruyuan North District in X Town>), the administrative jurisdiction is "M Sub-district" (<the entrance of Ruyuan North District in X Town, governedBy, M Sub-district>), and the time can be assumed to be "January 5, 2025, 08:00" (<Appeal 1, occursAt, 2025-01-05T08:00:00>). In the community governance knowledge graph, this information has been integrated with the governance-related data. For example, <Violating regulations, addressedBy, conduct investigation and governance>, <Violating regulations, addressedBy, dissuade and report to the administrative law enforcement unit>, <Comprehensive Law Enforcement Team, executes, conduct investigation and governance>, <Village Committee, executes, dissuade and report to the administrative law enforcement unit>, <Comprehensive Law Enforcement Team, operatesIn, M Sub-district>. These information are integrated through SPARQL query. For example: SELECT?solution?participant WHERE{?risk rdf:type ex:RegulationBreach; ex:locatedIn "the entrance of Ruyuan North District in X Town"; ex:governedBy "M Sub-district"; ex:addressedBy?solution.?participant ex:executes?solution; ex:operatesIn "M Sub-district"}, and the returned results are "conduct investigation and governance" and "Comprehensive Law Enforcement Team", "dissuade and report to the administrative law enforcement unit" and "Village Committee". When generating the governance plan, it is further optimized according to the query results combined with the decision-making requirements. The paragraph mentions that the "Comprehensive Law Enforcement Team" and the "Village Committee and Property Management Committee" are the cooperation networks in M Sub-district, and the knowledge graph may store <Comprehensive Law Enforcement Team, collaboratesWith, Village Committee>.For the risk of "violating regulations", the query results show two governance methods: If the risk level is high (such as serious traffic impact during peak hours), select "investigate, punish and manage", and the plan is "the comprehensive law enforcement team investigates, punishes and manages the illegal construction at the entrance of Ruyuan North District in X Town"; If the risk level is medium (such as non-peak hours), select "dissuade and report", and the plan is "the villagers' committee dissuades the illegal construction and reports to the administrative law enforcement unit". Assume that the appeal time is "08:00 on January 5, 2025" (traffic peak), and the risk level is "high". The final plan is "the comprehensive law enforcement team investigates, punishes and manages the illegal construction at the entrance of Ruyuan North District in X Town on January 5, 2025". The participating entity is the "comprehensive law enforcement team", and the collaboration network includes the "villagers' committee" providing assistance (such as on-site coordination). The plan is output in a structured form: <Plan 1, hasMeasure, investigate, punish and manage>, <Plan 1, hasParticipant, comprehensive law enforcement team>, <Plan 1, hasCollaborator, villagers' committee>.

[0084] The specific embodiments of this specification have been described above. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve the desired results. Additionally, the processes depicted in the figures do not necessarily require the particular order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0085] According to an embodiment of another aspect, a community governance decision-making generation device based on a knowledge graph is provided. Figure 8 A schematic block diagram of the community governance decision-making generation device based on the knowledge graph according to an embodiment is shown, as Figure 8 shown, the device 800 includes:

[0086] A data acquisition unit 801, configured to acquire community governance data, where the community governance data includes residents' appeals, geographical information, time information, and government department responsibility information.

[0087] A structured data generation unit 802, configured to perform information extraction and text processing on the community governance data using a data extraction model to generate structured data.

[0088] A sub-graph construction unit 803, configured to use the structured data to construct a causal analysis graph, a decision-making governance graph, and a spatio-temporal association graph; the causal analysis graph is used to describe the relationships among appeals, potential hazards, risks, events, and consequences; the decision-making governance graph is used to describe the relationships among appeals, governance participants, governance plans, and resource allocation; the spatio-temporal association graph is used to describe the relationships between appeals and time data, and between appeals and spatial data.

[0089] The atlas fusion unit 804 is configured to fuse the causal analysis atlas, the decision-making governance atlas, and the spatio-temporal association atlas to obtain a community governance knowledge atlas, and the community governance knowledge atlas supports semantic-based querying and reasoning.

[0090] The risk reasoning unit 805 is configured to obtain resident demand information, perform risk reasoning based on preset reasoning rules, and generate a risk reasoning result.

[0091] The solution generation unit 806 is configured to query in the community governance knowledge atlas based on the risk reasoning result and in combination with the spatio-temporal information defined by the spatio-temporal association atlas, and generate a community governance solution according to the query result.

[0092] As an implementable manner, the data collection unit 801 may be configured to include one or more of the following: implementing multi-source data integration through an API interface; cleaning, standardizing the format, and removing redundancy of the community governance data; performing data format conversion on the community governance data using an automated script, and removing low-confidence data using an anomaly detection algorithm.

[0093] As an implementable manner, when constructing the spatio-temporal association atlas, the sub-atlas construction unit 803 may be configured to: define the time attribute of the resident demand using the OWL-Time time ontology, and construct a time dimension atlas according to the time attribute, where the time dimension atlas includes the association relationship between the resident demand and the time information; define the spatial entity and its relationship using the GeoNames geographical ontology, and construct a spatial dimension atlas according to the spatial entity and its relationship, where the spatial dimension atlas includes the association relationship between the resident demand and the location information.

[0094] As an implementable manner, the data extraction model may be configured to: a fine-tuned UIE model and a fine-tuned BERT model; using the data extraction model to perform information extraction and text processing on the community governance data includes: using the fine-tuned UIE model to perform information extraction on the community governance data, and using the fine-tuned BERT model to perform text processing on the community governance data.

[0095] As an implementable approach, the graph fusion unit 804 can be configured to extract key entities and their relationships in the community governance scenario according to the causal analysis graph, the decision-making governance graph, and the spatio-temporal association graph, and construct an initial structure of the community governance knowledge graph; perform node embedding on the community governance knowledge graph based on a graph neural network, update node features using a graph convolutional network, and generate entity feature representations; adopt an attention mechanism to enhance the feature expression ability of associated entities, and perform weighted updates according to node similarity and semantic association; fine-tune the graph neural network in combination with the community governance prediction task, and output the dynamic association pattern between complex entities in the community governance knowledge graph.

[0096] As an implementable approach, the graph fusion unit 804 can be configured to obtain the causal relationship data defined in the causal analysis graph, where the causal relationship data includes semantic associations between residents' demands, hidden dangers, risks, events, and consequences; obtain the governance relationship data defined in the decision-making governance graph, where the governance relationship data includes collaboration rules between governance participants, governance plans, and resource allocation; obtain the spatio-temporal relationship data defined in the spatio-temporal association graph, where the spatio-temporal relationship data includes the time attributes and spatial attributes of community events based on the time ontology and geographical ontology; semantically align the causal relationship data, the governance relationship data, and the spatio-temporal relationship data, and establish cross-graph association rules, where the association rules include the mapping of causal relationships and spatio-temporal attributes, the connection of governance relationships and causal relationships, and the link of spatio-temporal attributes and governance relationships; based on the semantically aligned association rules, fuse and generate the community governance knowledge graph, and the community governance knowledge graph is represented in the form of entity nodes and relationship edges.

[0097] As an implementable approach, the solution generation unit 806 can be configured to determine the risk types and potential consequences related to residents' demands according to the risk reasoning result; use the time attributes defined in the spatio-temporal association graph to extract time information associated with the risk types from the knowledge graph, including the time point or time interval when the event occurs; use the spatial attributes defined in the spatio-temporal association graph to extract geographical location information associated with the risk types from the knowledge graph, including the community, block, or specific facility where the event occurs; based on the time information and geographical location information, construct spatio-temporal query conditions, perform semantic queries in the community governance knowledge graph, and obtain events, trends, or governance resources that match the risk types in the spatio-temporal dimension; according to the query results, combine the governance participants and resource allocation rules defined in the decision-making governance graph, and generate a community governance solution for the residents' demands, where the community governance solution includes governance measures and participating entities.

[0098] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other, and the differences between each embodiment and other embodiments are emphasized. In particular, for the device embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and for the relevant parts, reference can be made to the corresponding description in the method embodiments. The device embodiments described above are only illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative work.

[0099] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. The collection, use, and processing of the relevant data need to comply with the relevant laws, regulations, and standards of the relevant countries and regions, and corresponding operation entrances are provided for users to choose to authorize or refuse.

[0100] In addition, the embodiments of this application also provide a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps of the method described in any one of the foregoing method embodiments are implemented.

[0101] And an electronic device, including:

[0102] One or more processors; and a memory associated with the one or more processors, the memory is used to store program instructions, and when the program instructions are read and executed by the one or more processors, the steps of the method described in any one of the foregoing method embodiments are executed.

[0103] This application also provides a computer program product, including a computer program, and when the computer program is executed by a processor, the steps of the method described in any one of the foregoing method embodiments are implemented.

[0104] Among them, Figure 9Exemplarily shows the architecture of an electronic device, which may specifically include a processor 910, a video display adapter 911, a disk drive 912, an input / output interface 913, a network interface 914, and a memory 920. The above-mentioned processor 910, video display adapter 911, disk drive 912, input / output interface 913, network interface 914, and the memory 920 can be communicatively connected through a communication bus 930.

[0105] Among them, the processor 910 can be implemented in ways such as a general-purpose CPU, a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided by this application.

[0106] The memory 920 can be implemented in forms such as ROM (Read Only Memory), RAM (Random Access Memory), a static storage device, a dynamic storage device, etc. The memory 920 can store an operating system 921 for controlling the operation of the electronic device 900, and a basic input / output system (BIOS) 922 for controlling the low-level operations of the electronic device 900. In addition, it can also store a web browser 923, a data storage management system 924, and a community governance decision-making generation device 925 based on a knowledge graph, etc. The above-mentioned community governance decision-making generation device 925 based on a knowledge graph can be the application program that specifically implements the operations of the foregoing steps in the embodiments of this application. In short, when implementing the technical solutions provided by this application through software or firmware, the relevant program codes are stored in the memory X20 and are called and executed by the processor X10.

[0107] The input / output interface 913 is used to connect to an input / output module to achieve information input and output. The input / output module can be configured as a component in the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Among them, the input device can include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device can include a display, a speaker, a vibrator, an indicator light, etc.

[0108] The network interface 914 is used to connect to a communication module (not shown in the figure) to achieve communication and interaction between this device and other devices. Among them, the communication module can achieve communication through a wired method (such as USB, network cable, etc.) or through a wireless method (such as a mobile network, WIFI, Bluetooth, etc.).

[0109] The bus 930 includes a path for transmitting information among various components of the device, such as the processor 910, the video display adapter 911, the disk drive 912, the input / output interface 913, the network interface 914, and the memory 920).

[0110] It should be noted that although the above device only shows the processor 910, the video display adapter 911, the disk drive 912, the input / output interface 913, the network interface 914, the memory 920, the bus 930, etc., in the specific implementation process, the device may also include other components necessary for normal operation. In addition, those skilled in the art can understand that the above device may also only include the components necessary for implementing the solution of the present application, and does not necessarily include all the components shown in the figure.

[0111] From the description of the above embodiments, those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application, in essence or the part that contributes to the prior art, can be embodied in the form of a computer program product, which can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of the present application.

[0112] The above has introduced the technical solution provided by the present application in detail. Specific examples are used in this article to elaborate on the principle and implementation manner of the present application. The description of the above embodiments is only used to help understand the method and its core idea of the present application; at the same time, for those of ordinary skill in the art, according to the idea of the present application, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present application.

Claims

1. A method for generating community governance decisions based on a knowledge graph, characterized in that, The method includes: Collecting community governance data, which includes residents' demands, geographical information, time information, and government department responsibility information; Using a data extraction model to extract information and process the text of the community governance data to generate structured data; Using the structured data to construct a causal analysis graph, a decision-making governance graph, and a spatio-temporal association graph; the causal analysis graph is used to describe the relationships among demands, potential hazards, risks, events, and consequences; the decision-making governance graph is used to describe the relationships among demands, governance participants, governance solutions, and resource allocation; the spatio-temporal association graph is used to describe the relationships between demands and time data, and between demands and spatial data; Fusing the causal analysis graph, the decision-making governance graph, and the spatio-temporal association graph to obtain a community governance knowledge graph, which supports semantic-based query and reasoning; Obtaining residents' demand information, performing risk reasoning based on preset reasoning rules, and generating a risk reasoning result; Based on the risk reasoning result, combining with the spatio-temporal information defined by the spatio-temporal association graph, querying in the community governance knowledge graph, and generating a community governance solution according to the query result.

2. The method according to claim 1, wherein The collecting of the community governance data further includes one or more of the following: Implementing multi-source data integration through an API interface; Cleaning, standardizing the format, and removing redundancy from the community governance data; Using an automated script to perform data format conversion on the community governance data, and using an anomaly detection algorithm to remove low-confidence data.

3. The method according to claim 1, wherein The construction process of the spatio-temporal association graph includes: Using the OWL-Time time ontology to define the time attributes of the residents' demands, and constructing a time dimension graph according to the time attributes, where the time dimension graph includes the association relationship between residents' demands and time information; Using the GeoNames geographical ontology to define spatial entities and their relationships, and constructing a spatial dimension graph according to the spatial entities and their relationships, where the spatial dimension graph includes the association relationship between residents' demands and location information.

4. The method according to claim 1, characterized in that The data extraction model includes: a fine-tuned UIE model and a fine-tuned BERT model; the using of the data extraction model to extract information and process the text of the community governance data includes: Using the fine-tuned UIE model to extract information from the community governance data, and using the fine-tuned BERT model to process the text of the community governance data.

5. The method according to claim 1, wherein The fusing of the causal analysis graph, the decision-making governance graph, and the spatio-temporal association graph to generate a community governance knowledge graph includes: According to the causal analysis graph, the decision-making governance graph, and the spatio-temporal association graph, extracting key entities and their relationships in the community governance scenario, and constructing the initial structure of the community governance knowledge graph; Performing node embedding on the community governance knowledge graph based on a graph neural network, and using a graph convolutional network to update node features to generate entity feature representations; Using an attention mechanism to enhance the feature expression ability of associated entities, and performing weighted update according to node similarity and semantic association; Fine-tune the graph neural network in combination with the community governance prediction task, and output the dynamic association patterns between complex entities in the community governance knowledge graph.

6. The method according to claim 1, characterized in that, The fusion of the causal analysis graph, the decision-making governance graph, and the spatio-temporal association graph to obtain the community governance knowledge graph includes: Obtain the causal relationship data defined in the causal analysis graph, including the semantic associations between residents' demands, hidden dangers, risks, events, and consequences; Obtain the governance relationship data defined in the decision-making governance graph, including the collaboration rules between governance participants, governance plans, and resource allocation; Obtain the spatio-temporal relationship data defined in the spatio-temporal association graph, including the time attributes and spatial attributes of community events based on the time ontology and the geographical ontology; Semantically align the causal relationship data, the governance relationship data, and the spatio-temporal relationship data, and establish cross-graph association rules, where the association rules include the mapping of causal relationships and spatio-temporal attributes, the connection of governance relationships and causal relationships, and the link of spatio-temporal attributes and governance relationships; Based on the association rules after semantic alignment, fuse and generate the community governance knowledge graph, which is represented in the form of entity nodes and relationship edges.

7. The method according to claim 1, wherein The generation of the community governance plan based on the risk reasoning result, combined with the spatio-temporal information defined in the spatio-temporal association graph, and querying in the knowledge graph includes: According to the risk reasoning result, determine the risk types and potential consequences related to residents' demands; Utilize the time attributes defined in the spatio-temporal association graph to extract the time information associated with the risk type from the knowledge graph, including the time point or time interval when the event occurs; Utilize the spatial attributes defined in the spatio-temporal association graph to extract the geographical location information associated with the risk type from the knowledge graph, including the community, block, or specific facility where the event occurs; Based on the time information and geographical location information, construct spatio-temporal query conditions, perform semantic queries in the community governance knowledge graph, and obtain events, trends, or governance resources that match the risk type in the spatio-temporal dimension; According to the query result, combined with the governance participants and resource allocation rules defined in the decision-making governance graph, generate a community governance plan for the residents' demands, where the community governance plan includes governance measures and participating entities.

8. A method for generating community governance decisions based on a knowledge graph, characterized in that, The device includes: A data collection unit configured to collect community governance data, where the community governance data includes residents' demands, geographical information, time information, and government department responsibility information; A structured data generation unit configured to use a data extraction model to perform information extraction and text processing on the community governance data to generate structured data; A sub-graph construction unit configured to use the structured data to construct a causal analysis graph, a decision-making governance graph, and a spatio-temporal association graph; the causal analysis graph is used to describe the relationships between demands, hidden dangers, risks, events, and consequences; the decision-making governance graph is used to describe the relationships between demands, governance participants, governance plans, and resource allocation; the spatio-temporal association graph is used to describe the relationships between demands and time data, and demands and spatial data; The atlas fusion unit is configured to fuse the causal analysis atlas, the decision-making governance atlas, and the spatio-temporal association atlas to obtain a community governance knowledge atlas, and the community governance knowledge atlas supports semantic-based querying and reasoning; The risk reasoning unit is configured to obtain residents' appeal information, perform risk reasoning based on preset reasoning rules, and generate a risk reasoning result; The solution generation unit is configured to query in the community governance knowledge atlas based on the risk reasoning result and in combination with the spatio-temporal information defined by the spatio-temporal association atlas, and generate a community governance solution according to the query result.

9. An electronic device, characterized in that, Comprising: One or more processors; And A memory associated with the one or more processors, the memory being used to store program instructions, and when the program instructions are read and executed by the one or more processors, the steps of the method according to any one of claims 1 to 7 are executed.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images