A multi-attribute disaster emergency decision-making aid method based on knowledge graph

Abstract

The present application relates to a multi-attribute sudden disaster emergency decision-making assistance method based on a knowledge graph, comprising: constructing a knowledge graph; representing all historical sudden disaster emergency cases in the form of a knowledge graph, and each decision maker gives a different initial decision matrix according to the comprehensive situation of the historical sudden disaster emergency cases; based on the initial decision matrix, a normalized comprehensive influence matrix is ​​obtained through a TSK model; the normalized comprehensive influence matrix corresponding to each decision maker is used as the decision matrix given by each decision maker for a target case, and the decision matrix is ​​aggregated by the Louvain algorithm considering the comprehensive influence matrix to determine the generalized Shapley value of the fuzzy measure; the grey correlation between the target case and the historical sudden disaster emergency case is calculated based on the generalized Shapley value of the fuzzy measure, the historical similar cases are determined based on the grey correlation, and the target case is emergency handled according to the treatment measures in the historical similar cases.

Description

Technical Field

[0001] The present application relates to the field of emergency decision-making technology, and in particular to a multi-attribute sudden disaster emergency decision-making auxiliary method based on a knowledge graph. Background Art

[0002] After a sudden disaster occurs, there are characteristics of information asymmetry and information overload. The information obtained from different sources, information quality and accuracy are also uneven, resulting in decision-making lags and uncertainties. It is also difficult to find key information from massive amounts of information and data, analyze risks and take timely measures. In addition, in current emergency decision-making, there are problems such as difficulty in obtaining key attribute information, limited scope of crowd intelligence knowledge mining, and the traditional Louvain clustering algorithm, when dealing with large groups of experts in a social network environment, usually lacks a comprehensive consideration of the knowledge mastery, trust relationship and opinion similarity of decision-making experts.

[0003] Therefore, how to deal with the knowledge asymmetry and overload problems existing in these sudden disasters, filter key attribute information from massive knowledge, improve the level of group consensus while maintaining the autonomy of decision makers and reducing subsequent losses are issues that need to be addressed urgently. Summary of the invention

[0004] Based on this, it is necessary to provide a multi-attribute disaster emergency decision-making assistance method based on knowledge graph, which includes:

[0005] S1: Construct a knowledge graph based on the attributes of sudden disaster emergency cases and the relationships between sudden disaster emergency cases;

[0006] S2: All historical disaster emergency response cases are represented in the form of knowledge graphs, and each decision maker gives different initial decision matrices based on the comprehensive situation of historical disaster emergency response cases;

[0007] S3: Based on the initial decision matrix, the standardized comprehensive influence matrix is ​​obtained through the TSK model;

[0008] S4: taking the normalized comprehensive influence matrix corresponding to each decision maker as the decision matrix given by each decision maker for the target case, and aggregating the decision matrices by the Louvain algorithm considering the comprehensive influence matrix to determine the generalized Shapley value of the fuzzy measure;

[0009] S5: Calculate the grey correlation between the target case and historical emergency cases of sudden disasters based on the generalized Shapley value of fuzzy measurement, determine historical similar cases based on the grey correlation, and perform emergency treatment on the target case according to the treatment measures in the historical similar cases.

[0010] Beneficial effects: In this method, the decision maker gives an initial decision matrix based on historical cases, and applies the TSK model to construct a standardized comprehensive influence matrix. Then, the decision matrix is ​​aggregated by considering the Louvain algorithm improved by TSK. Next, the generalized Shapley value of the fuzzy measure is determined, and the fuzzy measure of the set composed of decision makers and attributes is comprehensively considered to reflect the overall contribution of each decision maker or set. Furthermore, the grey correlation coefficient and correlation degree between the target case and the historical emergency cases of sudden disasters are calculated, and corresponding treatment measures are provided for the target case with reference to the retrieved historical similar cases. This method comprehensively utilizes knowledge graphs, case reasoning and multi-attribute analysis to enhance the scientificity and practicality of decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0012] Figure 1 This is a flow chart of a multi-attribute sudden disaster emergency decision-making assistance method based on a knowledge graph in an embodiment of the present application. DETAILED DESCRIPTION

[0013] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are described in detail below in conjunction with the accompanying drawings. In the following description, many specific details are set forth to facilitate a full understanding of the present application. However, the present application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without violating the connotation of the present application, so the present application is not limited by the specific embodiments disclosed below.

[0014] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of this application, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0015] like Figure 1 As shown, this embodiment provides a multi-attribute sudden disaster emergency decision-making auxiliary method based on a knowledge graph, the method comprising:

[0016] S1: Construct a knowledge graph based on the attributes of sudden disaster emergency response cases and the relationships between sudden disaster emergency response cases.

[0017] Specifically, the construction process of the knowledge graph includes:

[0018] S1.1: Define the entity annotation symbols for each attribute in the sudden disaster emergency case;

[0019] S1.2: Use the BiLSTM+CRF model to sequence label the emergency cases of sudden disasters, and identify the entities in the emergency cases of sudden disasters based on the entity annotation symbols;

[0020] S1.3: Extract specific relationships between entities in emergency response cases;

[0021] S1.4: Connect two entities based on the specific relationship between the entities to form a triple data of <entity, relationship, entity>;

[0022] S1.5: Collect and store the triple data corresponding to each entity, and use the Neo4j tool to build the knowledge graph.

[0023] In this embodiment, the attributes include: event, location, time, cause, affected object, rescue measures, result, and follow-up action. The entity annotation symbols of the attributes are shown in Table 1;

[0024] Table 1 is a table of entity annotation symbols for attributes;

[0025] ;

[0026] In Table 1, B represents the beginning information of an entity, I represents the middle information of an entity, and O represents non-entity information; the suffixes EVE, LOC, TIM, CAU, AO, RM, OUT, and TUA represent events, locations, time, causes, affected objects, rescue measures, results, and follow-up actions, respectively.

[0027] In this embodiment, the specific relationships include: event-location, event-time, event-cause, event-affected object, event-rescue measures, event-result, event-follow-up action. These relationships are helpful to understand and analyze the structure, impact scope, response measures and other information of disaster events. The specific relationships are shown in Table 2;

[0028] Table 2 is a table of specific relationship type descriptions and examples;

[0029] ;

[0030] Furthermore, the BiLSTM layer in step S1.2 is a bidirectional LSTM, which can capture both forward and reverse information, making the use of text information more comprehensive and effective; the CRF layer is a conditional random field, which is a conditional distribution model for solving the output sequence under the condition of a given input sequence; the BiLSTM+CRF model provided in step S1.2 specifically includes the following steps:

[0031] Step 1: Symbolize the event description, for example, there is a n Sentences with words S , expressed as: ,in, Represents the first word in the sentence. Indicates the first n words;

[0032] Step 2: Each word The embedding layer maps the vector to a real number of fixed dimension. The output of the embedding layer is expressed as: , Indicates i The word vector representation of each word; represents the embedding layer; Indicates the first i words;

[0033] Step 3: Represent the word vector Input to BiLSTM, which consists of forward LSTM and backward LSTM, processing the forward and backward information of the input sequence respectively; concatenate the forward and backward hidden states to get the BiLSTM output of each word: ,in, Indicates i The BiLSTM output of words, Indicates i The forward hidden state of each word, Indicates i The backward hidden state of each word;

[0034] Step 4: Transform the output of the BiLSTM layer Mapped to a rating matrix through a fully connected layer, the dimension of the matrix is ​​( n , m ), m To mark the number of symbols for the entity, the output of the fully connected layer is expressed as: ,in, Indicates i Words belong to j The score of the entity annotation symbol;

[0035] Step 5: Input the scoring matrix into the CRF layer. The CRF layer corrects the predicted entity annotation symbols of BiLSTM by considering the transition probability between entity annotation symbols to ensure that the final label sequence is valid. The final label sequence is the entities identified in the sudden disaster emergency case.

[0036] In the knowledge graph, nodes represent entities. If there is a relationship between two nodes, they are connected by an undirected or directed edge, and this connection line is a specific relationship. The knowledge graph constructed in this way is not only convenient for information retrieval and analysis, but also can effectively support emergency decision-making and management.

[0037] S2: All historical disaster emergency response cases are represented in the form of knowledge graphs, and each decision maker gives a different initial decision matrix based on the comprehensive situation of historical disaster emergency response cases.

[0038] S3: Based on the initial decision matrix, the standardized comprehensive influence matrix is ​​obtained through the TSK model.

[0039] Specifically, the process of obtaining a standardized comprehensive influence matrix includes:

[0040] S3.1: Each decision maker scores the mutual trust level in the form of language terms to obtain a probabilistic language trust relationship matrix; the probabilistic language trust relationship matrix is ​​converted into a trust score matrix;

[0041] Further, the language terms used for scoring include:

[0042] The seven-granularity language term set is used for scoring, and the seven-granularity language term set is expressed as: s={s -3 = very low, s -2 = low, s -1 = slightly lower, s0 = average, s1 = slightly higher, s2 = high, s3 = very high}.

[0043] The trust score matrix is ​​expressed as:

[0044] ;

[0045] ;

[0046] ;

[0047] Where T represents the trust score matrix; Representing decision makers a For decision makers b The degree of trust; M represents the number of decision makers; Representing decision makers a For decision makers b Information measure of Represents the seventh granularity language terminology set i score level; Representation and The associated trust weight, for The complement of Represents a subscript function, Indicates the total number of language terms.

[0048] S3.2: Each decision maker scores each other's initial decision matrix in the form of language terms to obtain the individual decision score matrix of each decision maker; the individual decision score matrix is ​​expressed as: ,in, Indicates k The individual decision score matrix of decision makers, m represents the total number of solutions, n Indicates the total number of attributes, Indicated in the plan ,property Next k The decision score of each decision maker, Indicated in the plan ,property Next k A measure of information for a decision maker;

[0049] S3.3: Calculate the similarity between decision makers based on the individual decision score matrix of each decision maker and construct the similarity matrix S. ; The similarity calculation formula is:

[0050] ;

[0051] ;

[0052] in, Representing decision makers a With decision makers b The similarity between Indicated in the plan ,property Next decision maker a With decision makers b The similarity between Indicated in the plan ,property Next decision maker a The initial decision matrix; Indicated in the plan ,property Next decision maker b The initial decision matrix.

[0053] S3.4; Calculate the confidence of each decision maker based on information entropy; the confidence calculation formula is:

[0054] ;

[0055] ;

[0056] ;

[0057] ;

[0058] in, Indicates k The confidence level of each decision maker; represents information entropy; represents the set of normalized probabilistic language terms for decision maker k; represents the kth language term in the normalized probabilistic language term set; Representing language items probability; is the number of language items; is a constant. In order to ensure that the maximum information entropy is 1, this embodiment takes The value of is 1.28.

[0059] S3.5: Calculate the knowledge of each decision maker based on the average similarity between them and the confidence of each decision maker; the knowledge calculation formula is:

[0060] ;

[0061] in, Indicates k The knowledge level of each decision maker; represents the equilibrium parameter, ;if The larger the value, the more knowledgeable the decision makers are about the decision problem, and they will be very confident in giving an accurate evaluation.

[0062] S3.6: integrating the trust score matrix, the similarity matrix and the knowledge of each decision maker to obtain a comprehensive influence matrix;

[0063] The trust score reflects the degree of mutual trust between decision makers or subgroups. The similarity is used to measure the similarity of opinions between decisions. If two decision makers have a higher similarity, their opinions may be more consistent and comparable. The knowledge reflects the professional knowledge level of the decision maker or subgroup in a specific field or issue. The trust score matrix, the similarity matrix and the knowledge of each decision maker are integrated to obtain the comprehensive influence matrix; the expression of the comprehensive influence matrix is:

[0064] ;

[0065] Where CIM represents the comprehensive influence matrix; M represents the number of decision makers; represents the first weight; represents the second weight; ; Representing decision makers a With decision makers b Trust score between Representing decision makers a With decision makers b The similarity between represents the knowledge of the kth decision maker.

[0066] S3.7: Normalize the comprehensive influence matrix to obtain a normalized comprehensive influence matrix.

[0067] S4: The normalized comprehensive influence matrix corresponding to each decision maker is used as the decision matrix given by each decision maker for the target case, and the decision matrix is ​​assembled by the Louvain algorithm considering the comprehensive influence matrix to determine the generalized Shapley value of the fuzzy measure.

[0068] Specifically, the Louvain algorithm considering the comprehensive influence matrix includes:

[0069] Step 1: Assign each point in the decision matrix into an independent community;

[0070] Step 2: For each point, try to assign the point to other communities in turn, and calculate the modularity increment generated by adding the point to each community, record the community with the largest modularity increment, and when the maximum modularity increment is greater than zero, assign the point to the community with the largest modularity increment, otherwise it remains unchanged;

[0071] The calculation formula for modularity increment is:

[0072] ;

[0073] in, represents the modularity increment; Represents the sum of weight factors of all edges in the comprehensive influence matrix; Represents the sum of weight factors of the edges connected by the points in the comprehensive influence matrix; Indicate point The sum of the weight factors of the connected edges and other points in the comprehensive influence matrix;

[0074] Step 3: Repeat step 2 until the communities no longer change;

[0075] Step 4: Abstract all points belonging to the same community into a new point, and repeat steps 2-3 until the modularity of the entire comprehensive influence matrix no longer changes, and the algorithm ends;

[0076] The calculation formula of modularity is:

[0077] ;

[0078] Where Q represents modularity; N represents the number of decision makers; Representation and point The sum of the weight factors of all connected edges; Representation and point The sum of the weight factors of all connected edges; Indicate point The community you belong to; Indicate point The community you belong to; Indicate point With point The weight factor of the edge between them; represents the indicator function, when the point With point When they belong to the same community, ,otherwise .

[0079] Furthermore, the generalized Shapley value of the fuzzy measure is determined. In an actual decision, the decision maker usually gives the fuzzy measure subjectively, which will lead to randomness. n decision makers and n When there are two attributes, the fuzzy measures of both of them form 2 n The classical Choquet integral operator only uses 2 n The fuzzy measure is weighed by one of the sets, but in the process of dealing with practical problems, all sets need to be considered. Therefore, the fuzzy measure needs to be the average of a single decision maker, a single attribute, or a set consisting of a decision maker and an attribute.

[0080] S5: Calculate the grey correlation between the target case and historical emergency cases of sudden disasters based on the generalized Shapley value of fuzzy measurement, determine historical similar cases based on the grey correlation, and perform emergency treatment on the target case according to the treatment measures in the historical similar cases.

[0081] Specifically, the calculation formula of grey relational degree is:

[0082] ;

[0083] ;

[0084] ;

[0085] in, Indicates the target case and i The grey correlation between historical emergency disaster cases; Indicates that from j to n Attributes about fuzzy measure The generalized Shapley value of ; Indicates that from j to n A collection of attributes, Representing a collection The number of elements in; C represents the attribute set; Indicates that it is included in the set The set with the attribute set C, Representing a collection The number of elements in ; Indicates that from j+ 1 to n Attributes about fuzzy measure The generalized Shapley value of ; Indicates that from j+ 1 to n A collection of attributes; Indicates j The target case under the attribute i The grey correlation coefficient between historical emergency disaster cases; Representation attributes The target case Historical Disaster Response Cases The distance measure of Representation attributes The target case , Representation attributes Historical emergency response cases under the .

[0086] The historical similar cases determined based on the grey correlation degree include:

[0087] Based on the grey relational degree and similarity threshold, the calculation formula is:

[0088] ;

[0089] in, represents the similarity threshold; represents the degree coefficient, ; The larger the value of, the higher the similarity between the retrieved historical disaster emergency response cases and the target cases. The value of is set according to the actual situation; m Indicates the number of historical sudden disaster emergency response cases;

[0090] The grey correlation degree is compared with the similarity threshold, and when the grey correlation degree is greater than or equal to the similarity threshold, the corresponding historical disaster emergency response case is determined to be the historical similar case, and each historical similar case is formed into a historical similar case set;

[0091] Refer to the retrieved historical similar case set to take corresponding emergency response measures for the target case: First, comprehensively analyze the success and failure experiences in the historical similar case set, identify key decision points and effective strategies; second, adjust and optimize past strategies in combination with the specific context and needs of the target case to ensure applicability and timeliness; learn from the resource allocation and coordination mechanism in the historical similar case set to improve emergency response efficiency and resource utilization. During the implementation process, regularly monitor and evaluate the effectiveness of measures, adjust the response plan in a timely manner, and ensure that various measures are dynamically matched with actual needs. Through continuous learning and adaptation, enhance the flexibility and response speed of emergency response, and ultimately improve the overall emergency management effectiveness.

[0092] This embodiment uses case reasoning to assist decision making, which is achieved through the five core steps of case reasoning (case representation, case retrieval, case reuse, case correction and case retention), using past cases to help solve current problems. Case representation uses a knowledge graph method to represent cases with a graph data structure, with nodes representing entities or concepts and edges representing relationships, thereby establishing a historical case library. Case retrieval uses a matching algorithm to identify helpful past cases by screening out historical cases that are most relevant to the current problem from the historical case library. Case reuse applies the solution of similar historical cases to the current problem and adapts it to the current situation. Case correction modifies the solution according to the actual situation to ensure applicability. Case retention stores the solution process of new problems as new cases, enriches the case library, and provides reference for the future.

[0093] In the multi-attribute sudden disaster emergency decision-making auxiliary method based on knowledge graph provided in this embodiment, the decision maker gives an initial decision matrix based on historical cases, and applies the TSK model to construct a standardized comprehensive influence matrix. Then, the decision matrix is ​​assembled by considering the Louvain algorithm improved by TSK. Next, the generalized Shapley value of the fuzzy measure is determined, and the fuzzy measure of the decision maker and the attribute set is comprehensively considered to reflect the overall contribution of each decision maker or set. Further, the gray correlation coefficient and correlation degree between the target case and the historical sudden disaster emergency case are calculated, and corresponding treatment measures are provided for the target case with reference to the retrieved historical similar cases. This method comprehensively utilizes knowledge graphs, case reasoning and multi-attribute analysis to enhance the scientificity and practicality of decision-making.

[0094] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0095] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the patent application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent application shall be subject to the attached claims.

Claims

1. A multi-attribute disaster emergency decision-making auxiliary method based on knowledge graph, characterized in that: include: S1: Construct a knowledge graph based on the attributes of sudden disaster emergency cases and the relationships between sudden disaster emergency cases; S2: All historical disaster emergency response cases are represented in the form of knowledge graphs, and each decision maker gives different initial decision matrices based on the comprehensive situation of historical disaster emergency response cases; S3: Based on the initial decision matrix, the standardized comprehensive influence matrix is ​​obtained through the TSK model; The process of obtaining a standardized comprehensive influence matrix includes: S3.1: Each decision maker scores the mutual trust level in the form of language terms to obtain a probabilistic language trust relationship matrix; the probabilistic language trust relationship matrix is ​​converted into a trust score matrix; S3.2: Each decision maker scores each other's initial decision matrix in the form of language terms to obtain the individual decision score matrix of each decision maker; S3.3: Calculate the similarity between decision makers based on the individual decision score matrix of each decision maker and construct a similarity matrix; S3.4; Calculate the confidence of each decision maker based on information entropy; S3.5: Calculate the knowledge of each decision maker based on the average similarity between the decision makers and the confidence of each decision maker; S3.6: integrating the trust score matrix, the similarity matrix and the knowledge of each decision maker to obtain a comprehensive influence matrix; S3.7: Normalize the comprehensive influence matrix to obtain a normalized comprehensive influence matrix; S4: taking the normalized comprehensive influence matrix corresponding to each decision maker as the decision matrix given by each decision maker for the target case, and aggregating the decision matrices by the Louvain algorithm considering the comprehensive influence matrix to determine the generalized Shapley value of the fuzzy measure; The Louvain algorithm considering the comprehensive influence matrix includes: Step 1: Assign each point in the decision matrix into an independent community; Step 2: For each point, try to assign the point to other communities in turn, and calculate the modularity increment generated by adding the point to each community, record the community with the largest modularity increment, and when the maximum modularity increment is greater than zero, assign the point to the community with the largest modularity increment, otherwise it remains unchanged; Step 3: Repeat step 2 until the communities no longer change; Step 4: Abstract all points belonging to the same community into a new point, and repeat steps 2-3 until the modularity of the entire comprehensive influence matrix no longer changes, and the algorithm ends; S5: Calculate the grey correlation between the target case and historical emergency cases of sudden disasters based on the generalized Shapley value of fuzzy measurement, determine historical similar cases based on the grey correlation, and perform emergency treatment on the target case according to the treatment measures in the historical similar cases.

2. The multi-attribute disaster emergency decision-making auxiliary method based on knowledge graph according to claim 1 is characterized in that: In S1, the construction process of the knowledge graph includes: S1.1: Define the entity annotation symbols for each attribute in the sudden disaster emergency case; S1.2: Use the BiLSTM+CRF model to sequence label the emergency cases of sudden disasters, and identify the entities in the emergency cases of sudden disasters based on the entity annotation symbols; S1.3: Extract specific relationships between entities in emergency response cases; S1.4: Connect two entities based on the specific relationship between the entities to form a triple data of <entity, relationship, entity>; S1.5: Collect and store the triple data corresponding to each entity, and use the Neo4j tool to build the knowledge graph.

3. The multi-attribute sudden disaster emergency decision-making auxiliary method based on knowledge graph according to claim 2 is characterized in that: The attributes include: event, location, time, cause, affected object, rescue measures, result, and follow-up action.

4. The multi-attribute disaster emergency decision-making auxiliary method based on knowledge graph according to claim 3 is characterized in that: Specific relationships include: event-location, event-time, event-cause, event-affected object, event-rescue measures, event-result, and event-follow-up actions.

5. The multi-attribute disaster emergency decision-making auxiliary method based on knowledge graph according to claim 1 is characterized in that: The language terms used for scoring include: The seven-granularity language term set is used for scoring, and the seven-granularity language term set is expressed as: s={s -3 = very low, s -2 = low, s -1 = slightly lower, s0 = average, s1 = slightly higher, s2 = high, s3 = very high}.

6. The multi-attribute disaster emergency decision-making auxiliary method based on knowledge graph according to claim 1 is characterized in that: The expression of the comprehensive influence matrix is: ； Where CIM represents the comprehensive influence matrix; M represents the number of decision makers; represents the first weight; represents the second weight; Representing decision makers a With decision makers b Trust score between Representing decision makers a With decision makers b The similarity between represents the knowledge of the kth decision maker.

7. The multi-attribute disaster emergency decision-making auxiliary method based on knowledge graph according to claim 1 is characterized in that: The calculation formula of modularity is: ； Where Q represents modularity; N represents the number of decision makers; Representation and point The sum of the weight factors of all connected edges; Representation and point The sum of the weight factors of all connected edges; Indicate point The community you belong to; Indicate point The community you belong to; Indicate point With point The weight factor of the edge between them; Represents the indicator function, when the point With point When they belong to the same community, ,otherwise ; The calculation formula for modularity increment is: ； in, represents the modularity increment; Represents the sum of weight factors of all edges in the comprehensive influence matrix; Represents the sum of weight factors of the edges connected by the points in the comprehensive influence matrix; Indicate point The sum of the weight factors of the connected edges and other points in the comprehensive influence matrix.

8. The multi-attribute disaster emergency decision-making auxiliary method based on knowledge graph according to claim 1 is characterized in that: The calculation formula of grey relational degree is: ； ； ； in, Indicates the target case and i The grey correlation between historical emergency disaster cases; Indicates that from j to n Attributes about fuzzy measure The generalized Shapley value of ; Indicates that from j to n A collection of attributes, Representing a collection The number of elements in; C represents the attribute set; Indicates that it is included in the set The set with the attribute set C, Representing a collection The number of elements in ; Indicates that from j+ 1 to n Attributes about fuzzy measure The generalized Shapley value of ; Indicates that from j+ 1 to n A collection of attributes; Indicates j The target case under the attribute i The grey correlation coefficient between historical emergency disaster cases; Representation attributes The target case Historical Disaster Response Cases The distance measure of Representation attributes The target case , Representation attributes Historical emergency response cases under the .

Images