A method for recommending mobilization resources for maritime rescue missions

By analyzing maritime rescue missions using large-scale models and knowledge graph technology, and automatically recommending potential resources, the system solves the problems of slow response and inaccuracy in traditional methods, achieving efficient and rational resource utilization and improving the execution efficiency of maritime rescue.

CN121705504BActive Publication Date: 2026-11-17NO 30 INST OF CHINA ELECTRONIC TECH GRP CORP
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202511806446.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-12-03
Publication Date
2026-11-17
Estimated Expiration
2045-12-03

AI Technical Summary

Technical Problem

In traditional maritime rescue missions, resource recommendations rely on manual analysis and experience-based judgment, resulting in slow and inaccurate responses, making it difficult to meet the needs of rapid response and handle complex and ever-changing resource data and mission requirements.

Method used

By employing large-scale models, knowledge graphs, and word vector technologies, combined with a dedicated computational model, the system automatically analyzes maritime rescue mission descriptions, extracts key mission elements, calculates correlation coefficients for resource types, and generates a resource list through a potential resource scale calculation model.

Benefits of technology

It improves the response efficiency and resource utilization of maritime rescue missions, avoids material shortages, enhances rescue effectiveness, and provides reliable decision-making support for maritime rescue.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121705504B_ABST
    Figure CN121705504B_ABST
Patent Text Reader

Abstract

This application discloses a method for recommending mobilization resources for maritime rescue missions, comprising: user input of a maritime rescue mission description; analysis of the maritime rescue mission description based on a large model and extraction of key mission elements; combining maritime rescue knowledge graphs and word vector technology to obtain the types of resources that may need to be mobilized for the rescue mission, and calculating the correlation coefficient for each resource type; filtering potential resource types from the potential resource types for the rescue mission; using a built-in potential resource scale calculation model, obtaining an initial potential resource scale list based on key mission elements and potential resource types; comparing the initial potential resource scale list with the corresponding potential resources in the knowledge graph, and filtering from the initial potential resource scale list to obtain a final potential resource scale list. This application improves the accuracy, comprehensiveness, and scientific rigor of potential resource mobilization in maritime rescue scenarios.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of resource recommendation technology, and in particular to a method for recommending mobilization resources for maritime rescue missions. Background Technology

[0002] Maritime rescue missions, as a crucial component of mobilization efforts, hold unique strategic significance. These missions require the rapid deployment of various types of potential resources, including rescue vessels, helicopters, medical equipment, communication equipment, and specialized rescue personnel. Traditionally, the resources needed for maritime rescue missions rely heavily on manual analysis and experience-based judgment. However, this method suffers from several drawbacks: First, manual analysis is time-consuming, failing to meet the rapid response demands of emergency rescue missions and potentially missing optimal rescue opportunities. Second, experience-driven resource recommendations are susceptible to subjective influences, leading to incomplete or inaccurate recommendations and impacting the mission's effectiveness. Finally, traditional methods struggle to effectively handle massive amounts of resource data and complex, ever-changing mission requirements, especially in situations involving large-scale, multi-type resource coordination, where the limitations of traditional methods become even more pronounced. With the rapid development of artificial intelligence technology, how to utilize intelligent methods to quickly determine the necessary resources and improve the efficiency and effectiveness of maritime rescue missions has become a critical issue that urgently needs to be addressed. Summary of the Invention

[0003] In view of this, this application provides a method for recommending mobilization resources for maritime rescue missions, which improves the accuracy, comprehensiveness and scientific nature of potential resource allocation in maritime rescue scenarios.

[0004] This application discloses a method for recommending mobilization resources for maritime rescue missions, which includes: Step 1: The user inputs a description of the maritime rescue mission, which includes support requirements and specific mission content; Step 2: Analyze the maritime rescue mission description based on the large model and extract key mission elements; key mission elements include search and rescue scale, search and rescue time limit, search and rescue area, distance from shore and type of search and rescue target; Step 3: Combining the maritime rescue knowledge graph and word vector technology, the types of resources that may need to be mobilized for this rescue mission are obtained, and the correlation coefficient corresponding to each type of resource is calculated. The maritime rescue knowledge graph is used to store the association between maritime rescue missions and potential resource types, and the word vector technology is used to calculate the semantic similarity between key mission elements and potential resource types. The resource types include lifeboats, medical rescue boats, and oxygen bags. Step 4: Based on the correlation coefficient and preset threshold corresponding to each resource type, select potential resource types from the resource types that may need to be mobilized in this rescue mission; Step 5: Using the built-in potential resource scale calculation model, an initial potential resource scale list is obtained based on key task elements and potential resource types; the initial potential resource scale list includes resource type name, recommended scale and quality requirements; Step 6: Compare the initial potential resource size list with the corresponding potential resources in the knowledge graph, and filter from the initial potential resource size list to obtain the final potential resource size list.

[0005] Furthermore, the process of constructing the maritime rescue knowledge graph includes: Knowledge data is extracted from historical maritime rescue missions and their corresponding rescue plans. Entities, relationships, and attributes are extracted from the knowledge data, and knowledge merging, coreference resolution, and entity disambiguation are performed. The ontology is constructed based on the entities, relations, and attributes obtained from coreference resolution and entity disambiguation, and the strength of the relations between different entities is calculated. Based on the entities, relationships, and attributes obtained from coreference resolution and entity disambiguation, as well as the ontology and the strength of relationships between different entities, a maritime rescue knowledge graph is obtained.

[0006] Furthermore, the calculation of the relationship strength between different entities includes: Based on the probability of each entity appearing independently and the co-occurrence probability of each pair of entities, the historical co-occurrence probability of each pair of entities is obtained. The association strength between each pair of entities is obtained based on the historical co-occurrence probability of each pair of entities and the preset initial relationship strength; the preset initial relationship strength is a quantitative value of the association between two knowledge entities predefined by experts.

[0007] Furthermore, the step of obtaining the historical co-occurrence probability of each pair of entities based on the probability of each entity occurring independently and the co-occurrence probability of each pair of entities includes: The historical co-occurrence probability of every two entities can be obtained using the following formula:

[0008] in, Let x be the historical co-occurrence probability of entity x and knowledge entity y. Let x be the co-occurrence probability of knowledge entity x and knowledge entity y. and Let x and y be the probabilities of knowledge entities x and y appearing independently, respectively.

[0009] Furthermore, based on the historical co-occurrence probability of each pair of entities and the preset initial relationship strength, the association strength between each pair of entities is obtained, including: The strength of the association between any two entities can be obtained using the following formula:

[0010] in, The strength of the association between knowledge entity x and knowledge entity y. The historical co-occurrence probability weighting coefficient. Let x be the historical co-occurrence probability of entity x and knowledge entity y. Predefine the association weight coefficients for experts; The quantitative value of the association between knowledge entity x and knowledge entity y predefined by the expert, i.e., the initial relationship strength.

[0011] Further, step 3 includes: Step 31: Calculate the attenuation factor of the multi-hop relationship based on the number of path hops and the urgency of the current rescue mission; the attenuation factor of the multi-hop relationship is used to represent the degree to which the correlation coefficient gradually decreases as the relationship chain becomes longer; Step 32: Based on the decay factor of multi-hop relationships and the strength of association between different knowledge entities, obtain the correlation coefficient of resource types. The correlation coefficient refers to the degree of recommendation of potential resource types under the current mobilization task.

[0012] Further, step 31 includes: The attenuation factor for multi-hop relationships can be obtained using the following formula:

[0013] in, is the attenuation factor for multi-hop relationships; k is the path hop count, used to represent the distance from the root entity node; This represents the exponential decay coefficient; "state" indicates the urgency of the rescue mission, divided into emergency and normal times. is the linear decay coefficient; max is the maximum value function.

[0014] Further, step 32 includes: The correlation coefficient of resource types can be obtained using the following formula:

[0015] in, The correlation number represents the degree to which the mobilized resource type y is recommended under the current rescue mission. Representing two knowledge entities and The strength of the direct correlation between them.

[0016] Further, step 4 includes: Step 4: Based on the correlation coefficient and preset threshold, filter out the potential resource types from the types of resources that may need to be mobilized in this rescue mission; If the correlation coefficient for each resource type is greater than or equal to a preset threshold, then that resource type is recommended and designated as a potential resource type; otherwise, that resource type is not recommended.

[0017] Further, step 5 includes: For each type of potential resource, the built-in potential resource scale calculation model is invoked, and key task elements are input into the potential resource scale calculation model to obtain an initial potential resource scale list. The initial potential resource scale list includes the resource type name, recommended scale, and quality requirements.

[0018] Due to the adoption of the above technical solution, this application has the following advantages: 1. Compared to traditional methods relying on manual analysis and experience-based judgment, this application comprehensively utilizes intelligent technologies such as large-scale models, knowledge graphs, and word vector analysis, combined with a dedicated computational model, to significantly improve the response efficiency and resource utilization of maritime rescue missions. It effectively avoids problems such as material shortages and resource insufficiency during rescue operations, thereby enhancing rescue effectiveness. Overall, this method can provide reliable auxiliary decision-making support for maritime rescue-related operational agencies, improving the execution efficiency of maritime rescue work.

[0019] 2. By combining large models, knowledge graphs, and computational models, this application can effectively improve the intelligence level of potential resource recommendation in maritime rescue missions, and provide decision support for relevant agencies to carry out maritime rescue missions. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments recorded in the embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings.

[0021] Figure 1 This is a flowchart illustrating a method for recommending mobilization resources for maritime rescue missions, as described in an embodiment of this application. Figure 2 This is a flowchart illustrating the process of constructing a maritime rescue knowledge graph in this application embodiment; Figure 3 This is a flowchart illustrating the calculation of the association strength in the embodiments of this application; Figure 4 This is a flowchart illustrating the mobilization potential resource recommendation operation in the embodiments of this application. Figure 5(a) shows the actual operating effect in the embodiment of this application; Figure 5(b) shows the actual operating effect in the embodiment of this application. Detailed Implementation

[0022] The present application will be further described in conjunction with the accompanying drawings and embodiments. The described embodiments are only some, not all, of the embodiments of the present application. All other embodiments obtained by those skilled in the art should fall within the protection scope of the embodiments of the present application.

[0023] See Figure 1 This application provides an embodiment of a method for recommending mobilization resources for maritime rescue missions, which includes: Step 1: The user enters a description of the maritime rescue mission, which includes support requirements and specific mission details; Step 2: Analyze the maritime rescue mission description based on the large model and extract key mission elements; key mission elements include search and rescue scale, search and rescue time limit, search and rescue area, distance from shore and type of search and rescue target; Step 3: Combining the maritime rescue knowledge graph and word vector technology, we obtain the types of resources that may need to be mobilized for this rescue mission, and calculate the correlation coefficient for each type of resource. The maritime rescue knowledge graph is used to store the relationship between maritime rescue missions and potential resource types, and word vector technology is used to calculate the semantic similarity between key mission elements and potential resource types. Resource types include lifeboats, medical rescue boats, and oxygen bags. Step 4: Based on the correlation coefficient and preset threshold corresponding to each resource type, select potential resource types from the resource types that may need to be mobilized in this rescue mission; Step 5: Using the built-in potential resource scale calculation model, an initial potential resource scale list is obtained based on key task elements and potential resource types; the initial potential resource scale list includes resource type name, recommended scale and quality requirements; Step 6: Compare the initial potential resource size list with the corresponding potential resources in the knowledge graph, and filter from the initial potential resource size list to obtain the final potential resource size list.

[0024] In this embodiment of the application, the process of constructing a maritime rescue knowledge graph includes: Knowledge data is extracted from historical maritime rescue missions and their corresponding rescue plans. Entities, relationships, and attributes are extracted from the knowledge data, and knowledge merging, coreference resolution, and entity disambiguation are performed. The ontology is constructed based on the entities, relations, and attributes obtained from coreference resolution and entity disambiguation, and the strength of the relations between different entities is calculated. Based on the entities, relationships, and attributes obtained from coreference resolution and entity disambiguation, as well as the ontology and the strength of relationships between different entities, a maritime rescue knowledge graph is obtained.

[0025] In this embodiment of the application, calculating the strength of the relationship between different entities includes: Based on the probability of each entity appearing independently and the co-occurrence probability of each pair of entities, the historical co-occurrence probability of each pair of entities is obtained. The association strength between each pair of entities is obtained based on the historical co-occurrence probability of each pair of entities and the preset initial relationship strength; the preset initial relationship strength is a quantitative value of the association between two knowledge entities predefined by experts.

[0026] In this embodiment of the application, the historical co-occurrence probability of each pair of entities is obtained based on the probability of each entity occurring independently and the co-occurrence probability of each pair of entities, including: The historical co-occurrence probability of every two entities can be obtained using the following formula:

[0027] in, Let x be the historical co-occurrence probability of entity x and knowledge entity y. Let x be the co-occurrence probability of knowledge entity x and knowledge entity y. and Let x and y be the probabilities of knowledge entities x and y appearing independently, respectively.

[0028] In this embodiment of the application, the association strength between each pair of entities is obtained based on the historical co-occurrence probability of each pair of entities and a preset initial relationship strength, including: The strength of the association between any two entities can be obtained using the following formula:

[0029] in, The strength of the association between knowledge entity x and knowledge entity y. The historical co-occurrence probability weighting coefficient. Let x be the historical co-occurrence probability of entity x and knowledge entity y. Predefine the association weight coefficients for experts; The quantitative value of the association between knowledge entity x and knowledge entity y predefined by the expert, i.e., the initial relationship strength.

[0030] In this embodiment of the application, step 3 includes: Step 31: Calculate the attenuation factor of the multi-hop relationship based on the number of path hops and the urgency of the current rescue mission; the attenuation factor of the multi-hop relationship is used to represent the degree to which the correlation coefficient gradually decreases as the relationship chain becomes longer; Step 32: Based on the decay factor of multi-hop relationships and the strength of association between different knowledge entities, obtain the correlation coefficient of resource types. The correlation coefficient refers to the degree of recommendation of potential resource types under the current mobilization task.

[0031] In this embodiment of the application, step 31 includes: The attenuation factor for multi-hop relationships can be obtained using the following formula:

[0032] in, is the attenuation factor for multi-hop relationships; k is the path hop count, used to represent the distance from the root entity node; This represents the exponential decay coefficient; "state" indicates the urgency of the rescue mission, divided into emergency and normal times. is the linear decay coefficient; max is the maximum value function.

[0033] In this embodiment of the application, step 32 includes: The correlation coefficient of resource types can be obtained using the following formula:

[0034] in, The correlation number represents the degree to which the mobilized resource type y is recommended under the current rescue mission. Representing two knowledge entities and The strength of the direct correlation between them.

[0035] In this embodiment of the application, step 4 includes: Step 4: Based on the correlation coefficient and preset threshold, filter out the potential resource types from the types of resources that may need to be mobilized in this rescue mission; If the correlation coefficient for each resource type is greater than or equal to a preset threshold, then that resource type is recommended and designated as a potential resource type; otherwise, that resource type is not recommended.

[0036] In this embodiment of the application, step 5 includes: For each type of potential resource, the built-in potential resource scale calculation model is invoked, and key task elements are input into the potential resource scale calculation model to obtain an initial potential resource scale list. The initial potential resource scale list includes the resource type name, recommended scale, and quality requirements.

[0037] See Figure 2 The process of constructing a knowledge graph for maritime rescue.

[0038] In this process, knowledge extraction and fusion utilize industry-standard tools and methods, and therefore are not included in the protection scope of this method. The following section focuses on the specific process by which this method, based on extracted knowledge and a pre-built knowledge base, supplements the calculation of the strength of relationships between knowledge entities during knowledge processing. See [link to relevant documentation]. Figure 3 .

[0039] The pre-built knowledge base includes historical documents (including historical mobilization plans, emergency plans, and records of maritime rescue missions), professional terminology, and auxiliary information such as quantitative values ​​of the relationships between knowledge entities predefined by experts.

[0040] 1. Obtain knowledge entities Define the target knowledge entities x and y as "lifeboat" and "personnel" knowledge entities, respectively.

[0041] 2. Calculate the probability of historical sharing Based on historical mobilization plans, emergency response plans, and maritime rescue mission records, after analysis and calculation, the independent probabilities of knowledge entities x and y are 0.1 and 0.2, respectively, and their co-occurrence probability is 0.04898374. Calculate the historical co-occurrence probability of knowledge entities x and y using the following formula.

[0042]

[0043] In the formula: Let x be the historical co-occurrence probability of knowledge entity x and knowledge entity y; Let x be the co-occurrence probability of knowledge entity x and knowledge entity y. and Let x and y be the probabilities of knowledge entities x and y appearing independently, respectively.

[0044] The historical co-occurrence probability of the two knowledge entities, lifeboat (x) and personnel (y), is calculated to be 0.8956745.

[0045] 3. Calculate the strength of the association. The initial relationship strength between x and y is set to 0.8, and the historical co-occurrence probability weighting coefficient is... The correlation weight coefficient is 0.6. The value is 0.4. The strength of the correlation between x and y is calculated using the following formula:

[0046] In the formula: The strength of the association between knowledge entity x and knowledge entity y; These are the historical co-occurrence probability weighting coefficients, which can be flexibly adjusted through model parameter configuration according to actual conditions; Predefined association weight coefficients for experts can be flexibly adjusted through model parameter configuration according to actual conditions; The quantitative value of the association between knowledge entity x and knowledge entity y predefined by the expert, i.e., the initial relationship strength.

[0047] The calculated association strength between the two knowledge entities, lifeboat (x) and personnel (y), is 0.8574047. This value will be stored in the knowledge graph for future use.

[0048] See Figure 4 A method for recommending mobilization resources for maritime rescue missions, the specific technical solution of which is as follows: (1) The description of the maritime rescue mission is as follows: "Search for and rescue about 20 people who have fallen into the water within an area of ​​about 13 square kilometers 100 nautical miles offshore. The maritime search and rescue work needs to be completed within 8 hours. Organize nearby search and rescue forces as soon as possible to carry out the search and rescue, and prepare relevant first aid equipment and medical personnel." (2) The key task elements extracted after large-scale model analysis are shown in Table 1: Table 1 Key Task Element Information Table

[0049] (3) Based on the pre-constructed maritime rescue knowledge graph and combined with word vector technology, the potential resource types that may be involved in this mission are analyzed.

[0050] 1) Calculation of correlation coefficient of potential resources This embodiment uses a lifeboat as an example to describe the calculation process of the correlation coefficient of potential resources.

[0051] Assuming the current situation is an emergency, the exponential decay coefficient is... The direct correlation strength between the rescue vessel and "personnel" is 0.85361, the path hop count is 2, and the direct correlation strength between the rescue target "personnel" and the maritime rescue mission is 1. The calculation formula is as follows:

[0052]

[0053] In the formula: This is the decay factor for multi-hop relationships, used to represent the degree to which the correlation coefficient gradually decreases as the relationship chain lengthens. This represents the path hop count, indicating the distance from the root node. This is the exponential decay coefficient, used in emergency situations; The linear attenuation coefficient is used for normal scenarios; The correlation coefficient represents the degree of recommendation of the type of potential mobilization resource y in the current mobilization task. It indicates the strength of the direct association between two knowledge entities.

[0054] The correlation coefficient of the lifeboat was calculated to be 0.624747962177691.

[0055] In this embodiment, complete information on potential resource types derived from knowledge graph analysis is shown in Table 2: Table 2. Information on Potential Resource Types Based on Knowledge Graph Analysis

[0056] 2) Screening of potential resources Set recommended threshold parameters =0.3. Based on the screening rules, the types of potential resources after screening are shown in Table 3.

[0057] Table 3. Information on Potential Resource Types Based on Knowledge Graph Analysis

[0058] (4) Calculate the scale of potential resources For each of the aforementioned potential resource types, the corresponding built-in potential resource scale calculation model was invoked, inputting parameters such as search and rescue scale, search and rescue time limit, search and rescue area, distance from shore, and type of search and rescue target. Based on model analysis and calculation, the recommended scale and quality requirements for each type of potential resource are shown in Table 4. Table 4. Recommendation results calculated based on the built-in potential resource scale calculation model.

[0059] Note: "—" indicates that no corresponding recommended content was generated.

[0060] (5) Generate potential resource scale (including resource list screening) Based on the calculation results of the built-in potential resource scale calculation model, and through analysis and comparison with the attributes of the corresponding potential resources in the knowledge graph, it was found that the mobilization cycle of hospital ships, high-speed passenger ships, and medical rescue boats is greater than 8 hours, which does not meet the actual needs. After screening, the final potential resource scale list is shown in Table 5: Table 5. List of Potential Resource Scale

[0061] As demonstrated by the above implementation examples, this method can automatically recommend potential resources that may need to be mobilized based on the mission description of a maritime rescue mission, and can provide decision support for relevant organizations and personnel to carry out maritime rescue missions to a certain extent.

[0062] The actual operating results of this embodiment are shown in Figures 5(a) and 5(b).

[0063] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application and not to limit them. Although this application has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of this application. Any modifications or equivalent substitutions that do not depart from the spirit and scope of this application should be covered within the protection scope of the claims of this application.

Claims

1. A method for recommending mobilization resources for maritime rescue missions, characterized in that, include: Step 1: The user inputs a description of the maritime rescue mission, which includes support requirements and specific mission content; Step 2: Analyze the maritime rescue mission description based on the large model and extract key mission elements; key mission elements include search and rescue scale, search and rescue time limit, search and rescue area, distance from shore and type of search and rescue target; Step 3: Combining the maritime rescue knowledge graph and word vector technology, the types of resources that may need to be mobilized for this rescue mission are obtained, and the correlation coefficient corresponding to each type of resource is calculated. The maritime rescue knowledge graph is used to store the association between maritime rescue missions and potential resource types, and the word vector technology is used to calculate the semantic similarity between key mission elements and potential resource types. The resource types include lifeboats, medical rescue boats, and oxygen bags. Step 4: Based on the correlation coefficient and preset threshold corresponding to each resource type, select potential resource types from the resource types that may need to be mobilized in this rescue mission; Step 5: Using the built-in potential resource scale calculation model, an initial potential resource scale list is obtained based on key task elements and potential resource types; the initial potential resource scale list includes resource type name, recommended scale and quality requirements; Step 6: Compare the initial potential resource size list with the corresponding potential resources in the knowledge graph, and filter from the initial potential resource size list to obtain the final potential resource size list. Step 3 includes: Step 31: Calculate the attenuation factor of the multi-hop relationship based on the number of path hops and the urgency of the current rescue mission; the attenuation factor of the multi-hop relationship is used to represent the degree to which the correlation coefficient gradually decreases as the relationship chain becomes longer; Step 32: Based on the decay factor of multi-hop relationships and the strength of association between different knowledge entities, obtain the correlation coefficient of resource types. The correlation coefficient refers to the degree of recommendation of potential resource types under the current mobilization task. Step 31 includes: The attenuation factor for multi-hop relationships can be obtained using the following formula: in, is the attenuation factor for multi-hop relationships; k is the path hop count, used to represent the distance to the root entity node; This represents the exponential decay coefficient; "state" indicates the urgency of the rescue mission, divided into emergency and normal times. The linear decay coefficient is denoted by 'max', which is the maximum value function. Step 32 includes: The correlation coefficient of resource types can be obtained using the following formula: in, The correlation coefficient represents the degree to which the mobilized resource type y is recommended under the current rescue mission. Representing two knowledge entities and The strength of the direct correlation between them.

2. The method according to claim 1, characterized in that, The process of constructing the maritime rescue knowledge graph includes: Knowledge data is extracted from historical maritime rescue missions and their corresponding rescue plans. Entities, relationships, and attributes are extracted from the knowledge data, and knowledge merging, coreference resolution, and entity disambiguation are performed. The ontology is constructed based on the entities, relations, and attributes obtained from coreference resolution and entity disambiguation, and the strength of the relations between different entities is calculated. Based on the entities, relationships, and attributes obtained from coreference resolution and entity disambiguation, as well as the ontology and the strength of relationships between different entities, a maritime rescue knowledge graph is obtained.

3. The method according to claim 2, characterized in that, The calculation of the relationship strength between different entities includes: Based on the probability of each entity appearing independently and the co-occurrence probability of each pair of entities, the historical co-occurrence probability of each pair of entities is obtained. The association strength between each pair of entities is obtained based on the historical co-occurrence probability of each pair of entities and the preset initial relationship strength; the preset initial relationship strength is a quantitative value of the association between two knowledge entities predefined by experts.

4. The method according to claim 3, characterized in that, The process of obtaining the historical co-occurrence probability of each pair of entities based on the probability of each entity occurring independently and the co-occurrence probability of each pair of entities includes: The historical co-occurrence probability of every two entities can be obtained using the following formula: in, Let x be the historical co-occurrence probability of entity x and knowledge entity y. Let x be the co-occurrence probability of knowledge entity x and knowledge entity y. and Let x and y be the probabilities of knowledge entities x and y appearing independently, respectively.

5. The method according to claim 3 or 4, characterized in that, Based on the historical co-occurrence probability of each pair of entities and the preset initial relationship strength, the association strength between each pair of entities is obtained, including: The strength of the association between any two entities can be obtained using the following formula: in, The strength of the association between knowledge entity x and knowledge entity y. The historical co-occurrence probability weighting coefficient. Let x be the historical co-occurrence probability of entity x and knowledge entity y. Predefine the association weight coefficients for experts; The quantitative value of the association between knowledge entity x and knowledge entity y predefined by the expert, i.e., the initial relationship strength.

6. The method according to claim 1, characterized in that, Step 4 includes: Step 4: Based on the correlation coefficient and preset threshold, filter out the potential resource types from the types of resources that may need to be mobilized in this rescue mission; If the correlation coefficient for each resource type is greater than or equal to a preset threshold, then that resource type is recommended and designated as a potential resource type; otherwise, that resource type is not recommended.

7. The method according to claim 1, characterized in that, Step 5 includes: For each type of potential resource, the built-in potential resource scale calculation model is invoked, and key task elements are input into the potential resource scale calculation model to obtain an initial potential resource scale list. The initial potential resource scale list includes the resource type name, recommended scale, and quality requirements.

Citation Information

Patent Citations

  • Resource recommendation method and device and storage medium

    CN115982357A

  • Intelligent questioning and answering method and intelligent agent for large-model knowledge graph based on intention driving

    CN120851191A