A Recommended Method and Apparatus for Emergency Response to Marine Storm Surges

Abstract

This invention provides a method and apparatus for recommending emergency response plans for marine storm surges. The method includes: constructing a "scenario-response" disaster management system and a "scenario-response" visualized knowledge graph that integrates event chains and emergency plans. Through the constructed marine storm surge "scenario-response" disaster management system, "scenario-response" knowledge graph, and knowledge reasoning model, it can not only complete "scenario-response" matching of historical disaster events to achieve pre-disaster rehearsals and post-disaster analysis, but also perform "response" knowledge base text matching based on real-time generated disaster "scenario" text to generate emergency response plans and visualize the graph recommendation results, thereby achieving in-disaster response. This improves the entire process of disaster prevention and mitigation chain for marine storm surge disasters, enhances the generalization ability of knowledge reasoning methods, and provides important support for further improving the intelligent handling capabilities of marine storm surge disasters.

Description

Technical Field

[0001] This invention relates to the field of intelligent disaster response technology, and in particular to a method for recommending emergency response plans for marine storm surges, a device for recommending emergency response plans for marine storm surges, an electronic device, and a computer-readable medium. Background Technology

[0002] Current research methods in the field of intelligent disaster response mainly focus on disaster event chains and emergency plans. Research on disaster event chains further includes methods for constructing event chain knowledge and methods for disaster scenario simulation. Wang Ke et al. constructed six types of marine disaster chains, summarized the characteristics of marine disaster impacts, and established a comprehensive marine disaster chain, providing technical support for marine disaster prevention and control. Chen Xuelong et al. used Bayesian networks to analyze the attribute relationships between events to address the modeling problem of concurrent disaster event chains and used prior probabilities for event evolution analysis. Regarding storm surge disaster scenario construction and simulation methods, Rao Wenli et al. constructed typhoon storm surge disaster scenarios using dynamic Bayesian networks, calculated scenario state probabilities using prior and conditional probabilities, and realized key scenario simulations for typhoon storm surges. Zhang Chao et al. conducted research on scenario construction methods in the field of urban public safety, establishing complete risk event information through scenario construction, laying the foundation for subsequent risk analysis, risk assessment, and risk management. Wang Jiadong constructed typhoon storm surge event chains and conducted scenario simulations based on them. He also conducted chain-breaking disaster mitigation analysis from the perspective of protecting disaster-bearing bodies and proposed disaster mitigation strategies. Zhu Haiming et al. constructed a typhoon disaster chain knowledge graph based on the disaster risk census knowledge base, providing a new method and perspective for disaster chain research.

[0003] The "scenario-response" model stores and rehearses past disaster scenarios and response plans to address current disasters and potential secondary disasters. In recent years, many scholars have used the "scenario-response" model for rapid response to emergencies. Gong Qiansheng, focusing on emergency scenario identification, studied key technologies in "scenario-response" emergency decision-making, including the constituent elements of emergency scenario information, generation of scenario information needs, extraction and expression of objects within scenario elements, and scenario deduction. Zhang Gongxiao et al. established a hazardous chemical leak emergency decision-making system based on the "scenario-response" model and applied it to emergency rescue in chlorine leak accidents, improving emergency decision-making efficiency. Zhou Lin et al. used the "scenario-response" model to simulate flash flood disasters. Tang Zhaoping et al., through analysis and calculation of accident scenario sets, adopted the "scenario-response" model, considering the uncertainty of emergency resource needs after an emergency occurs, to improve the emergency response level of high-speed railways.

[0004] In summary, although there has been much research in the field of "scenario-response" intelligent disaster management, existing research on the construction methods of marine storm surge "scenario-response" models rarely combines event chains with emergency plan knowledge, and most of them are based on historical disaster scenarios and the response measures at that time, resulting in insufficient generalization ability and coarse-grained emergency response plans. Summary of the Invention

[0005] In view of the above problems, the present invention is proposed to provide a method for recommending emergency response plans for marine storm surges, a corresponding device for recommending emergency response plans for marine storm surges, an electronic device, and a computer-readable medium to overcome or at least partially solve the above problems.

[0006] This invention discloses a recommended method for emergency response to marine storm surges, the method comprising:

[0007] Construct a marine storm surge disaster response system; the marine storm surge disaster response system includes an event chain structure, a table of secondary derivative relationships of typical marine storm surge disaster events, a conceptual system of emergency plan knowledge elements, and an expression model of emergency plan entity relationships;

[0008] Based on the marine storm surge disaster response system, a visualized knowledge graph ontology for marine storm surge disaster response is generated. Based on this ontology, entities and relationships are jointly extracted from a large number of historical marine storm surge events and emergency plans to generate a visualized knowledge graph for marine storm surge disaster response.

[0009] Based on the event chain structure and the table of secondary derivative relationships of typical marine storm surge disaster events, the event chain of the current event is obtained by matching and deducing the current event; the event chain of the current event includes the disaster situation that has already occurred and the secondary derivative disasters that may occur in the future;

[0010] The event chain of the current event is matched with a visualized knowledge graph for handling marine storm surge disasters to obtain an emergency response plan for the current event. The emergency response plan for the current event is then visualized in the form of a knowledge graph and recommended to emergency decision-makers.

[0011] Optionally, a marine storm surge disaster response system may be established, including:

[0012] By analyzing the structural characteristics of a large number of historical marine storm surge events, the event chain structure was obtained. Then, by using the spatiotemporal constraint analysis method, the storm surge propagation mechanism and secondary derivation law of a large number of historical marine storm surge events were analyzed, and a table of secondary derivation relationships of typical marine storm surge disaster events was obtained.

[0013] A structured analysis of the emergency response plan for marine storm surges yielded a conceptual system of knowledge elements for the emergency response plan. The relationship between different types of elements was analyzed to derive an entity relationship expression model for the emergency response plan. The conceptual system of knowledge elements for the emergency response plan includes emergency response plan categories, emergency response plan procedures, and basic content of the emergency response plan.

[0014] Optionally, based on the marine storm surge disaster response system, a visualized knowledge graph ontology for marine storm surge disaster response is generated. Then, based on this ontology, entities and relationships are jointly extracted from a large number of historical marine storm surge events and emergency plans to generate a visualized knowledge graph for marine storm surge disaster response, including:

[0015] Based on the event chain structure and the secondary derivative relationship table of typical marine storm surge disaster events, the semantic expression formula of the disaster scenario ontology is generated by sorting out the conceptual hierarchy relationship, attribute relationship and association relationship in the disaster scenario. Based on the conceptual system of emergency plan knowledge elements and the entity relationship expression model of emergency plan, the semantic expression formula of the emergency response subject ontology is generated by sorting out the conceptual hierarchy relationship, attribute relationship and association relationship of the emergency plan. Combining the semantic expression formula of the disaster scenario ontology and the semantic expression formula of the emergency response subject ontology, a visual knowledge graph ontology of marine storm surge disaster response is generated.

[0016] Based on the ontology of a visualized knowledge graph for marine storm surge disaster response, the OSS-CasRel model is used to extract entities and relationships from a large number of historical marine storm surge events and emergency plans to generate a visualized knowledge graph for marine storm surge disaster response. The OSS-CasRel model is an improved CasRel model, which replaces the word segmentation dictionary of the original CasRel model with a self-built entity relationship corpus in the marine storm surge domain, and uses marine storm surge emergency plans for parameter training.

[0017] Optionally, the event chain of the current event is matched with a visualized knowledge graph for marine storm surge disaster management to obtain an emergency response plan for the current marine storm surge, including:

[0018] The text representations of event chains and emergency response plans in the visualized knowledge graph for marine storm surge disaster management are spliced ​​together, and the BERT-Base-Chinese model is used for text knowledge embedding and text vector feature extraction to obtain the event chain-embedded emergency response plan text embedding vector.

[0019] The BERT-Base-Chinese model is used to perform text knowledge embedding and text vector feature extraction on the text representation of the event chain of the current event, so as to obtain the text embedding vector of the event chain of the current event.

[0020] Calculate the matching degree between the text embedding vector of each event chain-emergency response plan and the text embedding vector of the event chain of the current event. Split the text embedding vector of the event chain-emergency response plan with the highest matching degree according to different attributes to obtain the marine storm surge emergency response plan for the current event.

[0021] Optionally, based on the event chain structure and the table of secondary derivative relationships of typical marine storm surge disaster events, the event chain of the current event is obtained by matching and deducing the current event, including:

[0022] Based on the event chain structure and the table of secondary derivative relationships of typical marine storm surge disaster events, the structural characteristics, storm surge propagation mechanism, and secondary derivative laws of the current event are analyzed to obtain the event chain of the current event.

[0023] This invention also discloses a device for recommending emergency response plans for marine storm surges, the device comprising:

[0024] The disaster response system construction module is used to build a marine storm surge disaster response system; the marine storm surge disaster response system includes an event chain structure, a table of secondary derivative relationships of typical marine storm surge disaster events, a conceptual system of emergency plan knowledge elements, and an emergency plan entity relationship expression model.

[0025] The disaster response visualization knowledge graph generation module is used to generate a marine storm surge disaster response visualization knowledge graph ontology based on the marine storm surge disaster response system. Based on the marine storm surge disaster response visualization knowledge graph ontology, it performs entity and relationship joint extraction on a large number of historical marine storm surge events and emergency plans to generate a marine storm surge disaster response visualization knowledge graph.

[0026] The real-time event chain matching and deduction module is used to match and deduce the event chain of the current event based on the event chain structure and the secondary derivative relationship table of typical marine storm surge disaster events. The event chain of the current event includes the disaster situation that has already occurred and the secondary derivative disasters that may occur in the future.

[0027] The emergency response plan generation module is used to match the event chain of the current event with the visualized knowledge graph of marine storm surge disaster management to obtain the marine storm surge emergency response plan for the current event, and to visualize and recommend the marine storm surge emergency response plan for the current event in the form of a knowledge graph to emergency decision-makers.

[0028] Optionally, the disaster response system construction module is also used for:

[0029] By analyzing the structural characteristics of a large number of historical marine storm surge events, the event chain structure was obtained. Then, by using the spatiotemporal constraint analysis method, the storm surge propagation mechanism and secondary derivation law of a large number of historical marine storm surge events were analyzed, and a table of secondary derivation relationships of typical marine storm surge disaster events was obtained.

[0030] A structured analysis of the emergency response plan for marine storm surges yielded a conceptual system of knowledge elements for the emergency response plan. The relationship between different types of elements was analyzed to derive an entity relationship expression model for the emergency response plan. The conceptual system of knowledge elements for the emergency response plan includes emergency response plan categories, emergency response plan procedures, and basic content of the emergency response plan.

[0031] Optionally, the disaster response visualization knowledge graph generation module is also used for:

[0032] Based on the event chain structure and the secondary derivative relationship table of typical marine storm surge disaster events, the semantic expression formula of the disaster scenario ontology is generated by sorting out the conceptual hierarchy relationship, attribute relationship and association relationship in the disaster scenario. Based on the conceptual system of emergency plan knowledge elements and the entity relationship expression model of emergency plan, the semantic expression formula of the emergency response subject ontology is generated by sorting out the conceptual hierarchy relationship, attribute relationship and association relationship of the emergency plan. Combining the semantic expression formula of the disaster scenario ontology and the semantic expression formula of the emergency response subject ontology, a visual knowledge graph ontology of marine storm surge disaster response is generated.

[0033] Based on the ontology of a visualized knowledge graph for marine storm surge disaster response, the OSS-CasRel model is used to extract entities and relationships from a large number of historical marine storm surge events and emergency plans to generate a visualized knowledge graph for marine storm surge disaster response. The OSS-CasRel model is an improved CasRel model, which replaces the word segmentation dictionary of the original CasRel model with a self-built entity relationship corpus in the marine storm surge domain, and uses marine storm surge emergency plans for parameter training.

[0034] Optionally, the emergency response plan generation module is further configured to:

[0035] The text representations of event chains and emergency response plans in the visualized knowledge graph for marine storm surge disaster management are spliced ​​together, and the BERT-Base-Chinese model is used for text knowledge embedding and text vector feature extraction to obtain the event chain-embedded emergency response plan text embedding vector.

[0036] The BERT-Base-Chinese model is used to perform text knowledge embedding and text vector feature extraction on the text representation of the event chain of the current event, so as to obtain the text embedding vector of the event chain of the current event.

[0037] Calculate the matching degree between the text embedding vector of each event chain-emergency response plan and the text embedding vector of the event chain of the current event. Split the text embedding vector of the event chain-emergency response plan with the highest matching degree according to different attributes to obtain the marine storm surge emergency response plan for the current event.

[0038] Optionally, the real-time event chain matching and deduction module is further used for:

[0039] Based on the event chain structure and the table of secondary derivative relationships of typical marine storm surge disaster events, the structural characteristics, storm surge propagation mechanism, and secondary derivative laws of the current event are analyzed to obtain the event chain of the current event.

[0040] The present invention also discloses an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0041] The memory is used to store computer programs;

[0042] When the processor executes the program stored in the memory, it implements the recommended method for emergency response to marine storm surges as described in this invention.

[0043] The present invention also discloses one or more computer-readable media having instructions stored thereon that, when executed by one or more processors, cause the processors to perform the recommended method for marine storm surge emergency response as described in the present invention.

[0044] This invention has the following advantages:

[0045] The method for recommending emergency response plans for marine storm surges of this invention constructs a "scenario-response" disaster management system and a "scenario-response" visualized knowledge graph that integrates event chains and emergency plans. Through the constructed marine storm surge "scenario-response" disaster management system, "scenario-response" knowledge graph, and knowledge reasoning model, it can not only complete "scenario-response" matching of historical disaster events, realizing pre-disaster rehearsals and post-disaster analysis, but also perform "response" knowledge base text matching based on real-time generated disaster "scenario" text, generating emergency response plans and visualizing the graph recommendation results to achieve in-disaster response. This improves the entire disaster prevention and mitigation chain for marine storm surge disaster events, enhances the generalization ability of knowledge reasoning methods, and provides important support for further improving the intelligent handling capabilities of marine storm surge disasters. Attached Figure Description

[0046] Figure 1 This is a flowchart illustrating the steps of a recommended method for an emergency response plan for marine storm surges provided in an embodiment of the present invention.

[0047] Figure 2 This is a schematic diagram of the entity relationship expression model for emergency response plans provided in an embodiment of the present invention;

[0048] Figure 3 This is the ontology design diagram of the marine storm surge "scenario-response" spatiotemporal knowledge graph provided in the embodiments of the present invention;

[0049] Figure 4 This is a diagram of the OSS-CasRel model architecture provided in this embodiment of the invention;

[0050] Figure 5 This is an overall flowchart of the "scenario-response" knowledge embedding and knowledge reasoning for marine storm surges provided in this embodiment of the invention;

[0051] Figure 6 This is a schematic diagram of the marine storm surge disaster response system provided in an embodiment of the present invention;

[0052] Figure 7 This is a schematic diagram of some nodes in the marine storm surge "scenario-response" knowledge graph provided in an embodiment of the present invention;

[0053] Figure 8 This is a schematic diagram illustrating examples of seawall breach disaster events and the "Longgang District Meteorological Disaster Emergency Plan" in the "Scenario-Response" knowledge graph provided in this embodiment of the invention;

[0054] Figure 9 This is a structural block diagram of a recommended device for an emergency response plan for marine storm surges provided in an embodiment of the present invention. Detailed Implementation

[0055] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0056] Reference Figure 1 The diagram illustrates a flowchart of a recommended method for an emergency response plan for marine storm surges provided in an embodiment of the present invention, which may specifically include the following steps:

[0057] Step 101: Construct a marine storm surge disaster response system; the marine storm surge disaster response system includes an event chain structure, a table of secondary derivative relationships of typical marine storm surge disaster events, a conceptual system of emergency plan knowledge elements, and an emergency plan entity relationship expression model;

[0058] Step 102: Based on the marine storm surge disaster response system, generate a visualized knowledge graph ontology for marine storm surge disaster response. Based on the visualized knowledge graph ontology for marine storm surge disaster response, perform joint entity and relationship extraction on a large number of historical marine storm surge events and emergency plans to generate a visualized knowledge graph for marine storm surge disaster response.

[0059] Step 103: Based on the event chain structure and the table of secondary derivative relationships of typical marine storm surge disaster events, match and deduce the event chain of the current event; the event chain of the current event includes the disaster situation that has already occurred and the secondary derivative disasters that may occur in the future;

[0060] Step 104: Match the event chain of the current event with the visualized knowledge graph for handling marine storm surge disasters to obtain the marine storm surge emergency response plan for the current event, and visualize and recommend the marine storm surge emergency response plan for the current event in the form of a knowledge graph to emergency decision-makers.

[0061] In a first embodiment of the present invention, a marine storm surge disaster response system is constructed, comprising:

[0062] By analyzing the structural characteristics of a large number of historical marine storm surge events, the event chain structure was obtained. Then, by using the spatiotemporal constraint analysis method, the storm surge propagation mechanism and secondary derivation law of a large number of historical marine storm surge events were analyzed, and a table of secondary derivation relationships of typical marine storm surge disaster events was obtained.

[0063] A structured analysis of the emergency response plan for marine storm surges yielded a conceptual system of knowledge elements for the emergency response plan. The relationship between different types of elements was analyzed to derive an entity relationship expression model for the emergency response plan. The conceptual system of knowledge elements for the emergency response plan includes emergency response plan categories, emergency response plan procedures, and basic content of the emergency response plan.

[0064] In a first embodiment of the present invention, a visualized knowledge graph ontology for marine storm surge disaster response is generated based on the marine storm surge disaster response system. Then, based on this ontology, entities and relationships are jointly extracted from a large number of historical marine storm surge events and emergency plans to generate a visualized knowledge graph for marine storm surge disaster response, including:

[0065] Based on the event chain structure and the secondary derivative relationship table of typical marine storm surge disaster events, the semantic expression formula of the disaster scenario ontology is generated by sorting out the conceptual hierarchy relationship, attribute relationship and association relationship in the disaster scenario. Based on the conceptual system of emergency plan knowledge elements and the entity relationship expression model of emergency plan, the semantic expression formula of the emergency response subject ontology is generated by sorting out the conceptual hierarchy relationship, attribute relationship and association relationship of the emergency plan. Combining the semantic expression formula of the disaster scenario ontology and the semantic expression formula of the emergency response subject ontology, a visual knowledge graph ontology of marine storm surge disaster response is generated.

[0066] Based on the ontology of a visualized knowledge graph for marine storm surge disaster response, the OSS-CasRel model is used to extract entities and relationships from a large number of historical marine storm surge events and emergency plans to generate a visualized knowledge graph for marine storm surge disaster response. The OSS-CasRel model is an improved CasRel model, which replaces the word segmentation dictionary of the original CasRel model with a self-built entity relationship corpus in the marine storm surge domain, and uses marine storm surge emergency plans for parameter training.

[0067] In a first embodiment of the present invention, the event chain of the current event is matched with a visualized knowledge graph for handling marine storm surge disasters to obtain an emergency response plan for the current marine storm surge, including:

[0068] The text representations of event chains and emergency response plans in the visualized knowledge graph for marine storm surge disaster management are spliced ​​together, and the BERT-Base-Chinese model is used for text knowledge embedding and text vector feature extraction to obtain the event chain-embedded emergency response plan text embedding vector.

[0069] The BERT-Base-Chinese model is used to perform text knowledge embedding and text vector feature extraction on the text representation of the event chain of the current event, so as to obtain the text embedding vector of the event chain of the current event.

[0070] Calculate the matching degree between the text embedding vector of each event chain-emergency response plan and the text embedding vector of the event chain of the current event. Split the text embedding vector of the event chain-emergency response plan with the highest matching degree according to different attributes to obtain the marine storm surge emergency response plan for the current event.

[0071] In a first embodiment of the present invention, the event chain of the current event is obtained by matching and deducing the event chain structure based on the event chain structure and the secondary derivative relationship table of typical marine storm surge disaster events, including:

[0072] Based on the event chain structure and the table of secondary derivative relationships of typical marine storm surge disaster events, the structural characteristics, storm surge propagation mechanism, and secondary derivative laws of the current event are analyzed to obtain the event chain of the current event.

[0073] 1. Method

[0074] Knowledge graph-based knowledge reasoning technology can establish connections between different entities in the "scenario-response" visualization process, integrating the knowledge expression needs of event chains with the visualization form of emergency plans. The implementation process of the "scenario-response" reasoning method integrating event chains is as follows: ① Construction of a "scenario-response" disaster response system integrating event chains and emergency plans. This involves analyzing the event chain construction and deduction process using historical disaster data; performing structured analysis of emergency plans related to marine storm surges; and constructing a "scenario-response" disaster response system. ② Construction of a "scenario-response" visualization knowledge graph. This involves defining the ontology layer of the "scenario-response" visualization knowledge graph, using entity and relation joint extraction technology to obtain the corresponding knowledge triples in the ontology layer, and constructing the "scenario-response" visualization knowledge graph. ③ "Scenario-response" knowledge embedding. This involves embedding knowledge into the knowledge graph and the data features of the disaster scenario input data, as well as the visualization dimension information of the emergency plan, converting high-dimensional corpus information into low-dimensional feature vectors. ④ Calculation of "scenario-response" matching degree. After obtaining the feature vector matrix, calculate the matching degree between the feature vector of the disaster scenario input data and the feature vector formed by the relevant nodes of the emergency response plan in the knowledge graph; ⑤ Visualize and recommend emergency response plans. Visualize and recommend emergency response plans and related content that have a high matching degree with the input data in the form of a knowledge graph to emergency decision-makers.

[0075] 1.1 Methodology for Constructing a "Scenario-Response" Disaster Response System Integrating Event Chains and Emergency Plans

[0076] Event chains, as a descriptive method for the evolution of disaster events, can clearly express the process by which a disaster event derives and secondary disasters in a temporal and spatial sequence. Emergency decision-makers can use event chains to match and extrapolate current events, and also to identify potential secondary disasters in the future, issuing early warnings. The construction of event chains typically involves extracting events through extensive historical case analysis, categorizing common event chain structures from different perspectives such as the causes of the event chain and the location of the sudden event, analyzing the structural characteristics of the event chain, and organizing the event nodes in a graphical form to reveal the evolutionary patterns between events.

[0077] In the structured analysis of emergency response plans, the analysis can be conducted from three levels: emergency response plan categories, emergency response plan procedures, and basic emergency response plan content, as shown in Table 1. The plan categories can be divided into provincial, municipal, district, street, unit, and grassroots organization levels according to hierarchy; and into general emergency response plans, special emergency response plans, and departmental emergency response plans according to type. By compiling relevant documents issued by national, provincial, and other levels of departments, such as the "National Natural Disaster Relief Emergency Response Plan," "Disaster Relief Emergency Work Procedures," and "Guangdong Provincial Meteorological Disaster Emergency Response Plan," and combining disaster risk management and emergency management theories, the emergency response plan procedures are divided into different stages according to the execution sequence of the disaster occurrence process: daily support stage, monitoring and early warning stage, emergency response stage, and post-disaster handling stage. Regarding the knowledge representation method of emergency response plans, an emergency knowledge element conceptual system is established using ontology theory, and the relationships between different types of elements are analyzed to derive an emergency response plan entity relationship expression model, as shown in Table 1. Figure 2 As shown.

[0078] Table 1: Summary of Emergency Response Plan Categories and Contents

[0079]

[0080] 1.2 Construction and Knowledge Embedding Methods of "Scenario-Response" Visualized Knowledge Graph

[0081] In the ontology design of the "scenario-response" knowledge graph for ocean storm surges, the "scenario" part is a unified description of the conceptual hierarchy, attribute relationships, and associations related to the disaster event chain and the disaster scenario itself. This invention represents a disaster scenario ontology semantically as follows:

[0082] DisasterScene={Scene_Concept, Scene_Property, Scene_Relation, Scene_Restriction, Scene_Instance}

[0083] Here, Scene_Concept represents the set of all disaster scenario concepts, including the definition of disaster scenario concepts and the classification of concept levels; Scene_Property represents the definition of the attributes of the disaster scenario itself, such as disaster name, time of occurrence, and location of occurrence; Scene_Relation represents the definition of the semantic relationship between disaster scenarios, such as secondary derivation, concurrency, and cluster occurrence; Scene_Restriction represents an axiom describing the constraint relationship between natural disaster scenarios, such as seawater intrusion causing floods, and seawater intrusion being part of the "rainstorm-flood" disaster chain; Scene_Instance represents a specific instance of a natural disaster scenario, such as a storm surge disaster occurring in Shenzhen, Guangdong Province.

[0084] The "Response" section of the marine storm surge "scenario-response" knowledge graph is a unified description of the conceptual hierarchy, attribute relationships, and correlations of emergency response plan data related to emergency forces, response content, and emergency command. An emergency response entity's ontology semantically represents:

[0085] ResponseSubject={Subject_Concept, Subject_Property, Subject_Relation, Subject_Restriction, Subject_Instance}

[0086] Among them, Subject_Concept represents the collection of all emergency response data concepts, including the definition of emergency response subject concepts and the classification of concept hierarchy; Subject_Property represents the definition of the attributes of the emergency response subject itself, such as the emergency response subject name, response subject type, etc.; Subject_Relation represents the definition of the semantic relationship between emergency response subjects, such as the inclusion relationship between subjects; Subject_Restriction represents an axiom, describing the constraint relationship between subjects; Subject_Instance represents a specific instance of an emergency response subject, such as a disaster emergency command center.

[0087] By combining the semantic expression formulas of "context" and "response," the ontology design of the "context-response" spatiotemporal knowledge graph for ocean storm surges is obtained, as follows: Figure 3 As shown, double arrows represent mapping rules, and single arrows represent inclusion relationships.

[0088] In previous research involving knowledge graph construction, entity and relation extraction has been the most tedious and time-consuming task. In recent years, various methods have been developed to accomplish this task. From the simplest but tedious manual extraction to rule-based extraction methods and then to deep learning model-based extraction methods, knowledge graph construction has become increasingly efficient. Among these, deep learning model-based extraction methods are relatively highly efficient and have become a popular approach in recent years.

[0089] Entity and relation joint extraction models can extract both entities and relations simultaneously, reducing the workload of aligning entities and relations and greatly improving the efficiency of graph construction. Among them, the CasRel model is well-known in the field of entity and relation joint extraction models and has high extraction accuracy. In particular, it solves the biggest challenge in entity and relation joint extraction—the problem of relation overlap.

[0090] This technical solution proposes an improved model, OSS-CasRel, based on the CasRel model. OSS is an abbreviation for marine storm surge, indicating that this model is an improved entity relation extraction model specifically for the marine storm surge domain. OSS-CasRel replaces the original word segmentation dictionary of the CasRel model, using a self-built entity relation corpus for the marine storm surge domain. The default word segmentation dictionary divides sentences by characters, meaning each character in a sentence is a word. This segmentation method loses the semantics of specific Chinese words. The self-built marine storm surge corpus contains most common geographical terms, which improves the model's accuracy in extracting entity relations in the marine storm surge domain, and also enhances the model's interpretability. The improved model is as follows: Figure 4 As shown. Secondly, the OSS-CasRel model was specifically trained with parameters for certain texts, such as emergency plans, which have a well-defined structure, making it more sensitive to entity relationships in such texts.

[0091] Knowledge embedding is a widely used knowledge representation method. Its main idea is to embed entities and relations from a knowledge graph into a continuous spatial vector. With the development of artificial intelligence technology, deep learning models have achieved remarkable results in natural language processing. For example, the BERT pre-trained deep language model has achieved significant performance improvements in many NLP tasks. BERT-Base-Chinese is a version of the BERT model specifically for Chinese text. It is pre-trained on Chinese text and can learn the semantic and syntactic information of Chinese text. Therefore, this invention uses the BERT-Base-Chinese model for text knowledge embedding and text vector feature extraction.

[0092] 1.3 Calculation of "Scenario-Response" Matching Degree and Knowledge Reasoning

[0093] This invention defines the "scenario-response" knowledge reasoning task as a matching task from "scenario" text to "response" text. That is, given a "scenario" text, the task finds the most matching "response" text from a knowledge graph as an emergency response plan. Here, the "scenario" text is represented as... ,in This represents the overall text of a disaster scenario. This represents the set of disaster events themselves and their related attribute nodes in the graph. This represents the set of event category nodes in the graph that are connected to the disaster event itself. This represents the set of event scale nodes in the graph that are connected to the disaster event itself. This represents the set of spatiotemporal attribute nodes in the graph that are connected to the disaster event itself. The text for "response" is represented as... ,in This text represents the overall text describing the response strategies under a specific disaster scenario. This represents the set of nodes containing the emergency response plan itself and its related attributes in the graph. This represents the set of response category nodes in the graph that are connected to the emergency plan itself. This represents the set of response level nodes in the graph that are connected to the emergency plan itself. This represents the set of spatiotemporal attribute nodes in the knowledge graph that are connected to the emergency response plan itself. Both the "scenario" text and the "response" text are composed of related nodes in the constructed marine storm surge "scenario-response" knowledge graph. This matching method based on the information of related nodes in the knowledge graph has stronger scientific validity and interpretability.

[0094] The overall process of knowledge embedding and reasoning in the "scenario-response" framework for ocean storm surges is as follows: Figure 5 As shown, firstly and Separated by semicolons, each character is considered a single unit. text The tokens are then used as input to the model, and after being segmented and vectorized into text by the tokenizer of the BERT-Base-Chinese model, the output is a token sequence. input_ids and used to determine the valid position in the token attention_mask The second step is to extract text features. First, load the pre-trained model, and then use the output from the previous step. input_ids and attention_mask The vector features of the text are calculated using these as parameters of the pre-trained model. last_hidden_state The third step is to calculate the text matching score. First, a downstream text matching task model is defined, and then the matching score between the "context" text and the "response" text is calculated using this model. out Finally, an emergency response plan is generated. First, the "response" text with the highest score after matching and calculating the current "scenario" is taken as the final result. Then, the "response" text with the highest score is split according to different attributes to generate the final emergency decision plan. Finally, the "response" reasoning results are visualized based on the constructed marine storm surge "scenario-response" knowledge graph.

[0095] 2 Experiments

[0096] 2.1 Construction of a "Scenario-Response" Disaster Response System Integrating Event Chains and Emergency Plans

[0097] In this invention, disaster event data is obtained from news reports, marine disaster bulletins, journals, and other sources over the past decade through a combination of web crawling and manual extraction. Then, based on the event chain construction and deductive analysis theory in Section 1.1, typical storm surge disaster events are identified, and spatiotemporal constraint analysis is used to analyze the storm surge propagation mechanism and secondary derivative laws. The final result is a table of secondary derivative relationships for typical marine storm surge disaster events, as shown in Table 2: disaster event levels are divided into four levels according to the secondary derivative laws.

[0098] Table 2 Secondary Derivative Relationships of Typical Marine Storm Surge Disasters

[0099]

[0100] The emergency response plans studied in this technical solution are mainly derived from the emergency response plans published on the websites of governments at the provincial, municipal, district, and street levels in Guangdong Province. After obtaining these plans through a combination of web scraping and manual extraction, plans related to marine storm surge disasters were further collected and organized, such as the "Guangdong Provincial Meteorological Disaster Emergency Response Plan," the "Shenzhen Municipal Flood Control Plan," the "Longgang District Meteorological Disaster Emergency Response Plan," and the "Henggang Street Natural Disaster Emergency Response Plan," totaling 1043 plans. Based on the structured analysis results of marine storm surge disaster events and emergency response plans presented above, the storm surge disaster response chain was analyzed, resulting in the marine storm surge disaster event response system as follows: Figure 6 As shown.

[0101] 2.2 Construction of a Visual Knowledge Graph for "Scenario-Response"

[0102] To fully leverage the response speed advantage of the "scenario-response" model and meet the needs of emergency personnel for response plans under different disaster scenarios, this invention proposes to recommend reasonable response plans to emergency personnel through a "scenario-response" visualized knowledge graph that integrates event chains and emergency plans. This invention first uses the disaster event caused by Typhoon Pearl (No. 1 of 2006) in Shenzhen, Guangdong Province, and the "Shenzhen Typhoon Emergency Plan" as examples. Following the ontology layer of the marine storm surge "scenario-response" visualized knowledge graph defined in Section 1.2, knowledge is extracted from the disaster event and emergency plan. The data for the knowledge extraction experiment comes from the storm surge disaster events and emergency plans collected above. The entities and relationships are organized into JSON format to form the training and validation sets of the model. The attribute names, meanings, and examples in the JSON data are shown in Table 3. The training set contains 13,414 data points, and the validation set contains 5,312 data points. After multiple experiments, the CasRel model achieved an accuracy of 65% on the validation set, and the OSS-CasRel model achieved an accuracy of 80% on the validation set. Experimental results show that the improved OSS-CasRel model can more accurately extract entities and relationships from marine storm surge disaster events and emergency response plan texts. Finally, the entity and relationship triples extracted by the OSS-CasRel model are stored in the neo4j database, completing the construction of a marine storm surge "scenario-response" knowledge graph. This case study extracted a total of 18,525 entities and 68 relationships. Figure 7 The diagram shows some of the nodes in the knowledge graph. Green nodes represent a specific event in the event chain; blue nodes represent emergency plans; purple nodes represent the command body; red nodes represent the response level; and gray nodes represent the handling body.

[0103] Table 3. Input data names, meanings, and examples for the OSS-CasRel model.

[0104]

[0105] 2.3 Experimental Methods and Case Analysis

[0106] To verify the accuracy of the marine storm surge "scenario-response" knowledge reasoning method, this invention conducts experiments using data from the disaster event caused by Typhoon Pearl in Shenzhen and the "Shenzhen Typhoon Emergency Response Plan" and "Longgang District Meteorological Disaster Emergency Response Plan." The overlap rate between the reasoning results and the original control experimental data for emergency information such as plan name, response entity, response content, command entity, plan category, plan level, response level, geographical location, response stage, and land use type is used as an evaluation index to macroscopically measure the accuracy of the recommended results. This section conducts "scenario-response" knowledge reasoning experiments based on the marine storm surge "scenario-response" knowledge graph constructed in Section 1.2 and the overall process of "scenario-response" matching degree calculation and knowledge reasoning introduced in Section 1.3. The experimental data for the "scenario-response" matching degree calculation model comes from the node information in the marine storm surge "scenario-response" knowledge graph, and the node information in the graph is integrated according to the expression formulas of the "scenario" text and the "response" text. Taking the breach of the Xichong coastal seawall in Dapeng Subdistrict, Longgang District, Shenzhen's eastern coast, and the "Longgang District Meteorological Disaster Emergency Plan" as examples, the schematic diagram of the representation of entities and relationships involved in its "scenario-response" model in the knowledge graph is as follows: Figure 8 The left half is shown. Green nodes represent disaster events; blue nodes represent emergency response plans; purple nodes represent the attribute categories of disaster events and emergency response plans; and red nodes represent specific attribute values. The assembled "scenario" text. "Response" text The knowledge graph node information involved is shown in Table 4. The land use type classification is based on the national standard "Classification of Current Land Use" (GB / T21010-2017), revised by a national-level department. The current "scenario" text is concatenated with each of the various "response" text inputs to form... Figure 5 The process, as shown, takes input and calculates the highest-scoring recommended "response" text for the current "scenario." Emergency information such as the emergency plan name, responsible entity, response content, command entity, plan category, plan level, response level, geographical location, response stage, and land use type are extracted. The original "response" text is then visualized within the "response" node information of the marine storm surge "scenario-response" knowledge graph. Figure 8 The right half shows: blue nodes represent the plan name; purple nodes represent the disaster event and the attribute category of the emergency plan; red nodes represent specific attribute values; brown nodes represent specific handling entities; and gray nodes represent specific handling content. Finally, the recommended "response" text output by the model is compared with the original "response" text in the "response" node information of the ocean storm surge "scenario-response" knowledge graph (as shown in the example in Table 4) to calculate the overlap rate between the inference result and the original data, and this is used as the accuracy of the model.

[0107] Table 4. "Scenario" and "Response" texts and their node information in the knowledge graph.

[0108]

[0109] 2.4 Overall Results and Analysis

[0110] The training set of the marine storm surge "scenario-response" knowledge reasoning model contains 1044 data points, and the validation set contains 642 data points. The "scenario" and "response" text data in the validation set are reasoned according to the knowledge embedding and knowledge reasoning process. The overlap rate between the reasoning results of the "response" text and the information of different nodes in the original visualization form of the experimental data is recorded. The experimental results are shown in Table 5.

[0111] Table 5 Overall Experimental Results

[0112]

[0113] Among them, the overlap rates for plan name, response entity, and plan level were 96%, 95%, and 95%, respectively, which are relatively high compared to other node attributes. This indicates that the model performs well in extracting features related to plan name, response entity, and plan level. This may be because the plan name, response entity, and plan level texts contain relatively distinct feature information, such as the affected area, disaster type, department name, and province / city / district name. In contrast, other node attributes, such as command entity, plan type, response level, geographical location, response stage, and land use, contain less distinctive feature information. Furthermore, the response content text is often lengthy and relational. The process is complex and difficult to extract, resulting in less than ideal matching results. However, in the actual use of the marine storm surge "scenario-response" knowledge graph by emergency responders, they only need to construct the current "scenario" text and input it into the model to obtain recommended "response" text. Based on the plan name information in the recommended "response" text, they can query the marine storm surge "scenario-response" knowledge graph for visualized graph results, including information on the emergency plan and related nodes such as the handling entity, handling content, command entity, plan category, plan level, response level, geographical location, handling stage, and land use type. From a practical application perspective, the generation of visualized graph recommendation results mainly depends on the overlap rate of plan names. Therefore, overall, the "scenario-response" reasoning model proposed in this invention has a high recommendation accuracy.

[0114] The present invention has the following advantages:

[0115] This invention integrates event chains and emergency plans to construct a "scenario-response" knowledge graph for marine storm surges. Through a "scenario-response" knowledge reasoning method, it enables rapid response to complex and ever-changing disaster "scenarios." Event chains clearly demonstrate the secondary and derivative relationships between disaster events, thus assisting emergency personnel in preventing potential secondary disasters. The attributes of each event within the event chain, such as the event itself, event category, event scale, and spatiotemporal attributes, constitute the "scenario" of the disaster event. In the knowledge graph, the attribute information of each disaster event is stored in the form of triple nodes, facilitating decision-makers to quickly form disaster "scenario" information based on the information of each node associated with the current disaster event. These "scenarios," linked through event chains, also form a "scenario chain." Emergency plans, as guiding documents for emergency response, are undeniably important in disaster prevention and mitigation. This invention, through analysis of the content and structure of emergency plans, completes the ontology design of emergency plans and decomposes the content of the emergency plans into triple nodes based on the ontology design, storing them in the knowledge graph, thereby improving the efficiency of emergency plan usage. In constructing the knowledge graph, this invention employs a joint entity and relation extraction model to complete the extraction work, and further improves the model's extraction effect by using a self-built corpus in the marine storm surge domain. Experimental results show that this approach can significantly improve the efficiency of knowledge graph construction and reduce manual costs.

[0116] To further improve the speed of generating emergency response plans for different disaster scenarios, this invention proposes a "scenario-response" knowledge reasoning method that integrates event chains and emergency plans. A deep learning model for text matching is used to achieve rapid matching of "scenario" text to "response" text. Experiments show that this method not only has high accuracy but also better interpretability due to its matching based on knowledge graph-related node information. In practical applications, disaster "scenarios" can be constructed using the disaster event "scenario" ontology design proposed in the technical solution. Then, the disaster "scenario" and various "response" texts are input into the model to calculate the matching degree. Finally, the "response" text with the highest matching degree is reverse-analyzed based on the emergency plan attribute information and combined with the related nodes in the knowledge graph to obtain a visualized emergency response plan.

[0117] Compared to previous technologies for emergency response to marine storm surge disasters, this invention firstly integrates event chains with emergency plan knowledge, visualizing knowledge relationships through a knowledge graph; secondly, it improves the efficiency and accuracy of knowledge reasoning by utilizing deep learning models and a self-built marine storm surge domain dictionary; thirdly, the model's input comes from the associated node information in the knowledge graph, which also enhances the model's interpretability; and finally, the marine storm surge "scenario-response" knowledge graph and knowledge reasoning model constructed by this invention can not only perform "scenario-response" matching for historical disaster events, enabling pre-disaster rehearsals and post-disaster analysis, but also match "response" knowledge base texts based on real-time generated disaster "scenario" texts, generating emergency response plans and visualizing the graph recommendation results to achieve in-disaster response. This improves the entire disaster prevention and mitigation chain for marine storm surge disasters, enhances the generalization ability of knowledge reasoning methods, and provides important support for further improving the intelligent response capabilities for marine storm surge disasters.

[0118] It should be noted that, for the sake of simplicity, the method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of the present invention are not limited to the described order of actions, because according to the embodiments of the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions involved are not necessarily essential to the embodiments of the present invention.

[0119] Reference Figure 9 The diagram shows a structural block diagram of a recommended device for emergency response to marine storm surges provided in an embodiment of the present invention, which may specifically include the following modules:

[0120] The disaster response system construction module 901 is used to construct a marine storm surge disaster response system; the marine storm surge disaster response system includes an event chain structure, a table of secondary derivative relationships of typical marine storm surge disaster events, a conceptual system of emergency plan knowledge elements, and an emergency plan entity relationship expression model.

[0121] The disaster response visualization knowledge graph generation module 902 is used to generate a marine storm surge disaster response visualization knowledge graph ontology based on the marine storm surge disaster response system, and to perform entity and relationship joint extraction on a large number of historical marine storm surge events and emergency plans based on the marine storm surge disaster response visualization knowledge graph ontology to generate a marine storm surge disaster response visualization knowledge graph.

[0122] The real-time event chain matching and deduction module 903 is used to match and deduce the event chain of the current event based on the event chain structure and the secondary derivative relationship table of typical marine storm surge disaster events. The event chain of the current event includes the disaster situation that has occurred and the secondary derivative disasters that may occur in the future.

[0123] The emergency response plan generation module 904 is used to match the event chain of the current event with the visualized knowledge graph of marine storm surge disaster management to obtain the marine storm surge emergency response plan for the current event, and to visualize and recommend the marine storm surge emergency response plan for the current event in the form of a knowledge graph to emergency decision-makers.

[0124] Optionally, the disaster response system construction module is also used for:

[0125] By analyzing the structural characteristics of a large number of historical marine storm surge events, the event chain structure was obtained. Then, by using the spatiotemporal constraint analysis method, the storm surge propagation mechanism and secondary derivation law of a large number of historical marine storm surge events were analyzed, and a table of secondary derivation relationships of typical marine storm surge disaster events was obtained.

[0126] A structured analysis of the emergency response plan for marine storm surges yielded a conceptual system of knowledge elements for the emergency response plan. The relationship between different types of elements was analyzed to derive an entity relationship expression model for the emergency response plan. The conceptual system of knowledge elements for the emergency response plan includes emergency response plan categories, emergency response plan procedures, and basic content of the emergency response plan.

[0127] Optionally, the disaster response visualization knowledge graph generation module is also used for:

[0128] Based on the event chain structure and the secondary derivative relationship table of typical marine storm surge disaster events, the semantic expression formula of the disaster scenario ontology is generated by sorting out the conceptual hierarchy relationship, attribute relationship and association relationship in the disaster scenario. Based on the conceptual system of emergency plan knowledge elements and the entity relationship expression model of emergency plan, the semantic expression formula of the emergency response subject ontology is generated by sorting out the conceptual hierarchy relationship, attribute relationship and association relationship of the emergency plan. Combining the semantic expression formula of the disaster scenario ontology and the semantic expression formula of the emergency response subject ontology, a visual knowledge graph ontology of marine storm surge disaster response is generated.

[0129] Based on the ontology of a visualized knowledge graph for marine storm surge disaster response, the OSS-CasRel model is used to extract entities and relationships from a large number of historical marine storm surge events and emergency plans to generate a visualized knowledge graph for marine storm surge disaster response. The OSS-CasRel model is an improved CasRel model, which replaces the word segmentation dictionary of the original CasRel model with a self-built entity relationship corpus in the marine storm surge domain, and uses marine storm surge emergency plans for parameter training.

[0130] Optionally, the emergency response plan generation module is further configured to:

[0131] The text representations of event chains and emergency response plans in the visualized knowledge graph for marine storm surge disaster management are spliced ​​together, and the BERT-Base-Chinese model is used for text knowledge embedding and text vector feature extraction to obtain the event chain-embedded emergency response plan text embedding vector.

[0132] The BERT-Base-Chinese model is used to perform text knowledge embedding and text vector feature extraction on the text representation of the event chain of the current event, so as to obtain the text embedding vector of the event chain of the current event.

[0133] Calculate the matching degree between the text embedding vector of each event chain-emergency response plan and the text embedding vector of the event chain of the current event. Split the text embedding vector of the event chain-emergency response plan with the highest matching degree according to different attributes to obtain the marine storm surge emergency response plan for the current event.

[0134] Optionally, the real-time event chain matching and deduction module is further used for:

[0135] Based on the event chain structure and the table of secondary derivative relationships of typical marine storm surge disaster events, the structural characteristics, storm surge propagation mechanism, and secondary derivative laws of the current event are analyzed to obtain the event chain of the current event.

[0136] As the device embodiment is basically similar to the method embodiment, the description is relatively simple, and relevant parts can be found in the description of the method embodiment.

[0137] In addition, embodiments of the present invention also provide an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus.

[0138] Memory, used to store computer programs;

[0139] When the processor executes the program stored in the memory, it implements the recommended method for emergency response to marine storm surges as described in the above embodiments.

[0140] The communication bus mentioned above can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not mean that there is only one bus or one type of bus.

[0141] The communication interface is used for communication between the aforementioned terminal and other devices.

[0142] The memory may include random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0143] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0144] In another embodiment of the present invention, a computer-readable storage medium is also provided, which stores instructions that, when executed on a computer, cause the computer to perform the recommended method for emergency response to marine storm surges as described in the above embodiments.

[0145] In another embodiment of the present invention, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute the recommended method for emergency response to marine storm surges described in the above embodiments.

[0146] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially as a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0147] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0148] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the apparatus embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0149] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A recommended method for emergency response to marine storm surges, characterized in that, The method includes: Construct a marine storm surge disaster response system; the marine storm surge disaster response system includes an event chain structure, a table of secondary derivative relationships of typical marine storm surge disaster events, a conceptual system of emergency plan knowledge elements, and an expression model of emergency plan entity relationships; Based on the marine storm surge disaster response system, a visualized knowledge graph ontology for marine storm surge disaster response is generated. Based on this ontology, entities and relationships are jointly extracted from a large number of historical marine storm surge events and emergency plans to generate a visualized knowledge graph for marine storm surge disaster response. Based on the event chain structure and the table of secondary derivative relationships of typical marine storm surge disaster events, the event chain of the current event is obtained by matching and deducing the current event; the event chain of the current event includes the disaster situation that has already occurred and the secondary derivative disasters that may occur in the future; The event chain of the current event is matched with a visualized knowledge graph for handling marine storm surge disasters to obtain an emergency response plan for the current event. The emergency response plan for the current event is then visualized in the form of a knowledge graph and recommended to emergency decision-makers. Based on the marine storm surge disaster response system, a visualized knowledge graph ontology for marine storm surge disaster response is generated. Then, based on this ontology, entities and relationships are jointly extracted from a large number of historical marine storm surge events and emergency plans to generate a visualized knowledge graph for marine storm surge disaster response, including: Based on the event chain structure and the secondary derivative relationship table of typical marine storm surge disaster events, the semantic expression formula of the disaster scenario ontology is generated by sorting out the conceptual hierarchy relationship, attribute relationship and association relationship in the disaster scenario. Based on the conceptual system of emergency plan knowledge elements and the entity relationship expression model of emergency plan, the semantic expression formula of the emergency response subject ontology is generated by sorting out the conceptual hierarchy relationship, attribute relationship and association relationship of the emergency plan. Combining the semantic expression formula of the disaster scenario ontology and the semantic expression formula of the emergency response subject ontology, a visual knowledge graph ontology of marine storm surge disaster response is generated. Based on the ontology of a visualized knowledge graph for marine storm surge disaster response, the OSS-CasRel model is used to extract entities and relationships from a large number of historical marine storm surge events and emergency plans to generate a visualized knowledge graph for marine storm surge disaster response. The OSS-CasRel model is an improved CasRel model, which replaces the word segmentation dictionary of the original CasRel model with a self-built entity relationship corpus in the marine storm surge domain, and uses marine storm surge emergency plans for parameter training.

2. The method according to claim 1, characterized in that, Establish a marine storm surge disaster response system, including: By analyzing the structural characteristics of a large number of historical marine storm surge events, the event chain structure was obtained. Then, by using the spatiotemporal constraint analysis method, the storm surge propagation mechanism and secondary derivation law of a large number of historical marine storm surge events were analyzed, and a table of secondary derivation relationships of typical marine storm surge disaster events was obtained. A structured analysis of the emergency response plan for marine storm surges yielded a conceptual system of knowledge elements for the emergency response plan. The relationship between different types of elements was analyzed to derive an entity relationship expression model for the emergency response plan. The conceptual system of knowledge elements for the emergency response plan includes emergency response plan categories, emergency response plan procedures, and basic content of the emergency response plan.

3. The method according to claim 1, characterized in that, By matching the event chain of the current event with a visualized knowledge graph for marine storm surge disaster management, an emergency response plan for the current marine storm surge event is obtained, including: The text representations of event chains and emergency response plans in the visualized knowledge graph for marine storm surge disaster management are spliced ​​together, and the BERT-Base-Chinese model is used for text knowledge embedding and text vector feature extraction to obtain the event chain-embedded emergency response plan text embedding vector. The BERT-Base-Chinese model is used to perform text knowledge embedding and text vector feature extraction on the text representation of the event chain of the current event, so as to obtain the text embedding vector of the event chain of the current event. Calculate the matching degree between the text embedding vector of each event chain-emergency response plan and the text embedding vector of the event chain of the current event. Split the text embedding vector of the event chain-emergency response plan with the highest matching degree according to different attributes to obtain the marine storm surge emergency response plan for the current event.

4. The method according to claim 1, characterized in that, Based on the event chain structure and the table of secondary derivative relationships of typical marine storm surge disaster events, the event chain of the current event is obtained by matching and deducing the current event, including: Based on the event chain structure and the table of secondary derivative relationships of typical marine storm surge disaster events, the structural characteristics, storm surge propagation mechanism, and secondary derivative laws of the current event are analyzed to obtain the event chain of the current event.

5. A device for recommending emergency response plans for marine storm surges, characterized in that, The device includes: The disaster response system construction module is used to build a marine storm surge disaster response system; the marine storm surge disaster response system includes an event chain structure, a table of secondary derivative relationships of typical marine storm surge disaster events, a conceptual system of emergency plan knowledge elements, and an emergency plan entity relationship expression model. The disaster response visualization knowledge graph generation module is used to generate a marine storm surge disaster response visualization knowledge graph ontology based on the marine storm surge disaster response system. Based on the marine storm surge disaster response visualization knowledge graph ontology, it performs entity and relationship joint extraction on a large number of historical marine storm surge events and emergency plans to generate a marine storm surge disaster response visualization knowledge graph. The real-time event chain matching and deduction module is used to match and deduce the event chain of the current event based on the event chain structure and the secondary derivative relationship table of typical marine storm surge disaster events. The event chain of the current event includes the disaster situation that has already occurred and the secondary derivative disasters that may occur in the future. The emergency response plan generation module is used to match the event chain of the current event with the visualized knowledge graph of marine storm surge disaster management to obtain the marine storm surge emergency response plan for the current event, and to visualize and recommend the marine storm surge emergency response plan for the current event in the form of a knowledge graph to emergency decision-makers. The disaster response visualization knowledge graph generation module is also used for: Based on the event chain structure and the secondary derivative relationship table of typical marine storm surge disaster events, the semantic expression formula of the disaster scenario ontology is generated by sorting out the conceptual hierarchy relationship, attribute relationship and association relationship in the disaster scenario. Based on the conceptual system of emergency plan knowledge elements and the entity relationship expression model of emergency plan, the semantic expression formula of the emergency response subject ontology is generated by sorting out the conceptual hierarchy relationship, attribute relationship and association relationship of the emergency plan. Combining the semantic expression formula of the disaster scenario ontology and the semantic expression formula of the emergency response subject ontology, a visual knowledge graph ontology of marine storm surge disaster response is generated. Based on the ontology of a visualized knowledge graph for marine storm surge disaster response, the OSS-CasRel model is used to extract entities and relationships from a large number of historical marine storm surge events and emergency plans to generate a visualized knowledge graph for marine storm surge disaster response. The OSS-CasRel model is an improved CasRel model, which replaces the word segmentation dictionary of the original CasRel model with a self-built entity relationship corpus in the marine storm surge domain, and uses marine storm surge emergency plans for parameter training.

6. The apparatus according to claim 5, characterized in that, The disaster response system construction module is also used for: By analyzing the structural characteristics of a large number of historical marine storm surge events, the event chain structure was obtained. Then, by using the spatiotemporal constraint analysis method, the storm surge propagation mechanism and secondary derivation law of a large number of historical marine storm surge events were analyzed, and a table of secondary derivation relationships of typical marine storm surge disaster events was obtained. A structured analysis of the emergency response plan for marine storm surges yielded a conceptual system of knowledge elements for the emergency response plan. The relationship between different types of elements was analyzed to derive an entity relationship expression model for the emergency response plan. The conceptual system of knowledge elements for the emergency response plan includes emergency response plan categories, emergency response plan procedures, and basic content of the emergency response plan.

7. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; The memory is used to store computer programs; When the processor executes the program stored in the memory, it implements the recommended method for emergency response to marine storm surges as described in any one of claims 1-4.

8. One or more computer-readable media having instructions stored thereon that, when executed by one or more processors, cause the processors to perform the recommended method of the marine storm surge emergency response scheme as described in any one of claims 1-4.

Images