A method and device for reconstructing a communication network based on historical communication results

Abstract

The present invention provides a method and device for reconstructing a propagation network based on historical propagation results, including constructing an infection status data propagation network, obtaining a parent-child relationship evaluation index between any two nodes and then combining a soft clustering method to obtain a propagation topology matrix set, performing a genetic algorithm operation to obtain a new propagation probability matrix and a corresponding new propagation topology matrix, establishing an adaptability function, obtaining a first adaptability function value of a propagation topology matrix and a corresponding propagation probability matrix, and obtaining a second adaptability function value of a new propagation topology matrix and a corresponding new propagation probability matrix, and replacing the propagation topology matrix and the propagation probability matrix according to the first adaptability function value and the second adaptability function value. The present invention can more accurately infer the influence relationship between nodes in a propagation network and reconstruct a potential propagation network.

Description

Technical Field

[0001] The present invention relates to the field of information dissemination technology, and in particular to a method and device for reconstructing a dissemination network based on historical dissemination results. Background Art

[0002] In today's information society, the speed and breadth of information dissemination have reached an unprecedented level. From hot events on social media to the spread of epidemics in the field of public health, the phenomenon of dissemination is ubiquitous and has far-reaching impacts. As an abstract model that describes the path of information dissemination and the mutual influence relationship between nodes, the dissemination network has shown great application value in many fields. For example, in the field of public opinion analysis, the information dissemination network can be analyzed to understand the context of public opinion changes and provide solid support for public opinion supervision and regulation; in the field of public health, the disease transmission network can be analyzed to predict the trend of the epidemic and scientifically plan prevention and control measures. However, in practical applications, it is often relatively difficult to directly obtain the mutual influence relationship between nodes in the dissemination network. What is easy to obtain is the dissemination data that is easy to observe. The reconstruction of the dissemination network is based on these historical dissemination data to explore the potential influence relationship in the network, in order to provide effective guidance for future dissemination activities.

[0003] Current research on propagation network reconstruction is mainly based on the exact infection time data of nodes in the historical propagation process, and reveals the mutual influence between nodes through the order of infection. However, in real situations, due to the complexity of the propagation process, accurately capturing the node infection time requires a lot of resources to continuously and carefully monitor the propagation process, and the cost of such monitoring is often high and unaffordable. Although some methods can reconstruct the propagation network based on the infection status data of the nodes, the parent-child influence relationship evaluation indicators they use often cannot well characterize the parent-child relationship between nodes, and the hard clustering strategy adopted has misjudgments on the boundary conditions of the influence relationship evaluation indicators, affecting the accuracy of the reconstruction results. Summary of the invention

[0004] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a method and device for reconstructing a propagation network based on historical propagation results. This method not only abandons the dependence on accurate infection time information, but also uses a soft clustering method to obtain preliminary results of propagation network reconstruction according to the proposed parent-child influence relationship evaluation index, and uses a genetic algorithm for further optimization, so as to more accurately infer the influence relationship between nodes in the propagation network and reconstruct the potential propagation network.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0006] A method for reconstructing a propagation network based on historical propagation results comprises the following steps:

[0007] Construct an infection status data propagation network and obtain the parent-child relationship evaluation index between any two nodes;

[0008] Based on the parent-child relationship evaluation index between any two nodes and the soft clustering method, a propagation topology matrix set is obtained;

[0009] Construct a corresponding propagation probability matrix for each propagation topology matrix;

[0010] Perform genetic algorithm operation to obtain a new propagation probability matrix and a corresponding new propagation topology matrix;

[0011] Establishing an adaptability function, obtaining a first adaptability function value of a propagation topology matrix and a corresponding propagation probability matrix, and obtaining a second adaptability function value of a new propagation topology matrix and a corresponding new propagation probability matrix;

[0012] The difference between the first fitness function value and the second fitness function value is obtained. If the difference is greater than 0, the propagation topology matrix is ​​replaced by the new propagation topology matrix and the propagation probability matrix is ​​replaced by the new propagation probability matrix. Otherwise, the propagation topology matrix and the propagation probability matrix remain unchanged.

[0013] Furthermore, the calculation formula for the parent-child relationship evaluation index between any two nodes based on the propagation network is:

[0014]

[0015] in:

[0016]

[0017] 1≤i,j≤n

[0018] In the formula, PRI(v j →v i ) represents node v j is node v i The possibility of the parent node, n is the number of nodes in the propagation network, and They represent the node v after the lth propagation process ends. i and node v j The infection state value, β represents the number of propagation processes; I(·) is an indicator function, which takes the value of 1 when the condition in the brackets is met and takes the value of 0 when it is not met; γ ab are predefined weight parameters, a∈[0,1], b∈[0,1].

[0019] Furthermore, the method for obtaining the propagation topology matrix set based on the parent-child relationship evaluation index between any two nodes combined with the soft clustering method is:

[0020] The fuzzy C-means clustering algorithm with a cluster number of 2 is used to cluster all parent-child relationship evaluation indicators to obtain the membership matrix U, where Represents node v i and node v j The membership degree between them is obtained by clustering the evaluation index of the parent-child relationship. Indicates the degree of membership of the category with a smaller value of the parent-child relationship evaluation index. represents the degree of membership of the category with a larger value of the parent-child relationship evaluation index, and and Satisfaction between n is the number of nodes in the propagation network;

[0021] The membership matrix is ​​sampled several times to obtain a set of propagation topology matrices.

[0022] Furthermore, the calculation formula for constructing the corresponding propagation probability matrix for each propagation topology matrix is:

[0023]

[0024] In the formula, A m is the propagation topology matrix, W m is the propagation probability matrix, (A m ) i,j Indicates A m The value of the element in the i-th row and j-th column; rand(0,1) represents a random number between 0 and 1; (W m ) i,j W m The value of the element in the i-th row and j-th column, n is the number of nodes in the propagation network.

[0025] Furthermore, the method of performing genetic algorithm operation to obtain a new propagation probability matrix and a corresponding new propagation topology matrix is:

[0026] Perform mutation operation on each propagation probability matrix to obtain the first updated propagation probability matrix, and the calculation formula is:

[0027]

[0028] Where W m is the propagation probability matrix, represents the first updated propagation probability matrix The element in the i-th row and j-th column in ; F is a preset scaling factor; (W m ) r1,r2 Represents the matrix W m The element in the r1th row and r2th column of , and its value satisfies r1≠i,r2≠j; (W m ) r3,r4Represents the matrix W m The element in the r3th row and r4th column of , and its value satisfies r3≠i,r3≠r1,r4≠j,r4≠r2; (W m ) r5,r6 Represents the matrix W m The element in the r5th row and r6th column of , and the value satisfies r5≠i,r5≠r1,r5≠r3,r6≠j,r6≠r2,r6≠r4, where n is the number of nodes in the propagation network;

[0029] The second updated propagation probability matrix is ​​obtained by performing a crossover operation on each first updated propagation probability matrix, and the calculation formula is:

[0030]

[0031] In the formula, represents the second updated propagation probability matrix The element in the i-th row and j-th column; CR is a preset crossover probability parameter; rand(0,1) represents a random number generated between 0 and 1;

[0032] An update operation is performed on each second updated propagation probability matrix to obtain a third updated propagation probability matrix, and the calculation formula is:

[0033]

[0034]

[0035] In the formula, represents the second updated propagation probability matrix The element in the i-th row and j-th column of , μ is a preset parameter; Indicates that after the lth propagation process ends, node v j The infection state value of , β represents the number of propagation processes; (A m ) i,j Indicates A m The value of the element in row i and column j, A m represents the propagation topology matrix; Indicates that after the lth propagation process ends, node v t The infection status value of (A m ) t,j Indicates A m The value of the element in row t and column j, represents the second updated propagation probability matrix The element at row t and column j in ;

[0036] Using the third updated propagation probability matrix as a new propagation probability matrix;

[0037] Based on the new propagation probability matrix, the new propagation topology matrix is ​​obtained. The calculation formula is:

[0038]

[0039]

[0040] Where μ is a preset parameter.

[0041] Furthermore, the calculation formula of the fitness function is:

[0042]

[0043] In the formula, A m is the propagation topology matrix, W m is the propagation probability matrix, λ is a pre-set parameter; (W m ) j,i W m The element in the jth row and ith column of ; and They represent the node v after the lth propagation process ends. i and node v j The infection status value of m ) i,j Indicates A m The value of the element in row i and column j, (A m ) j,i Indicates A m The value of the element in the jth row and the ith column, β represents the number of propagation processes, n is the number of nodes in the propagation network, 1≤i,j≤n.

[0044] Furthermore, all differences are traversed, and if the maximum value of the difference is less than a given threshold τ, the iteration is stopped, otherwise the genetic algorithm operation is returned to obtain a new propagation probability matrix and a corresponding new propagation topology matrix.

[0045] A communication network reconstruction device based on historical communication results, comprising:

[0046] The propagation network construction module is used to construct the infection status data propagation network and obtain the parent-child relationship evaluation index between any two nodes;

[0047] The propagation topology matrix acquisition module is used to obtain a propagation topology matrix set based on the parent-child relationship evaluation index between any two nodes combined with the soft clustering method;

[0048] A propagation probability matrix acquisition module is used to construct a corresponding propagation probability matrix for each propagation topology matrix;

[0049] An updating module, used for performing genetic algorithm operation to obtain a new propagation probability matrix and a corresponding new propagation topology matrix;

[0050] An adaptability function building module, used to establish an adaptability function, obtain a first adaptability function value of a propagation topology matrix and a corresponding propagation probability matrix, and obtain a second adaptability function value of a new propagation topology matrix and a corresponding new propagation probability matrix;

[0051] A replacement module is used to obtain the difference between the first fitness function value and the second fitness function value. If the difference is greater than 0, the propagation topology matrix is ​​replaced by the new propagation topology matrix and the propagation probability matrix is ​​replaced by the new propagation probability matrix. Otherwise, the propagation topology matrix and the propagation probability matrix remain unchanged.

[0052] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the method for reconstructing a propagation network based on historical propagation results is implemented.

[0053] A non-transitory computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned method for reconstructing a propagation network based on historical propagation results.

[0054] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0055] The present invention first uses soft clustering combined with sampling strategies to construct multiple initial propagation networks, then iterates and updates the propagation network through a genetic algorithm, and finally finds the optimal propagation network through a fitness function. The present invention not only abandons the reliance on precise infection time information, but also uses a soft clustering method to obtain preliminary results of propagation network reconstruction based on the proposed parent-child influence relationship evaluation index, and uses a genetic algorithm for further optimization, so as to more accurately infer the influence relationship between nodes in the propagation network, reconstruct the potential propagation network, reduce the boundary misjudgment of the influence relationship evaluation index, and increase the accuracy of the reconstruction results. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The illustrative embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0057] Figure 1 A flow chart of a method for reconstructing a propagation network based on historical propagation results according to the present invention;

[0058] Figure 2 The result of the corresponding F value of the propagation network reconstructed on the artificial network generated by the specific implementation method of the present invention;

[0059] Figure 3It is a schematic diagram of a propagation network reconstruction device based on historical propagation results of the present invention. DETAILED DESCRIPTION

[0060] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0061] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and the like are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first", "second", and the like may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.

[0062] Example 1

[0063] Embodiment 1 provides a method for reconstructing a propagation network based on historical propagation results, such as Figure 1 As shown, the following steps are included:

[0064] Step S1: construct an infection status data propagation network and obtain the parent-child relationship evaluation index between any two nodes;

[0065] Step S2: Based on the parent-child relationship evaluation index between any two nodes and the soft clustering method, a propagation topology matrix set is obtained;

[0066] Step S3: construct a corresponding propagation probability matrix for each propagation topology matrix;

[0067] Step S4: performing a genetic algorithm operation to obtain a new propagation probability matrix and a corresponding new propagation topology matrix;

[0068] Step S5: Establishing an adaptability function, obtaining a first adaptability function value of a propagation topology matrix and a corresponding propagation probability matrix, and obtaining a second adaptability function value of a new propagation topology matrix and a corresponding new propagation probability matrix;

[0069] Step S6: Obtain the difference between the first fitness function value and the second fitness function value. If the difference is greater than 0, replace the propagation topology matrix with the new propagation topology matrix and replace the propagation probability matrix with the new propagation probability matrix. Otherwise, the propagation topology matrix and the propagation probability matrix remain unchanged.

[0070] This embodiment provides a method for reconstructing a propagation network based on historical propagation results. First, a soft clustering combined with a sampling strategy is used to construct multiple initial propagation networks. Then, the propagation network is iteratively updated through a genetic algorithm. Finally, the optimal propagation network is found through a fitness function. The present invention not only abandons the reliance on accurate infection time information, but also uses a soft clustering method to obtain preliminary results of propagation network reconstruction based on the proposed parent-child influence relationship evaluation index, and uses a genetic algorithm for further optimization, so as to more accurately infer the influence relationship between nodes in the propagation network, reconstruct the potential propagation network, reduce the boundary misjudgment of the influence relationship evaluation index, and increase the accuracy of the reconstruction results.

[0071] In step S1 of this embodiment, the method for constructing the infection status data propagation network is:

[0072] Use a directed graph G = {V, A, W} to represent the propagation network;

[0073] V represents the set of nodes in the communication network {v 1 ,v 2 ,...,v n}, n is the number of nodes in the propagation network;

[0074] A is an n×n matrix representing the topological structure of the communication network. i,j ∈{0,1} represents node v i To node v j Is there a directed edge between A i,j =0 means there is no slave node v i To node v j A i,j =1 indicates that there is a slave node v i To node v j The directed edge of node v i For node v j The parent node of

[0075] W represents the n×n matrix of the propagation probability corresponding to the directed edge, W i,j Indicates that when node v i While infected, they will i,j The probability of infecting the uninfected node v j ;

[0076] use represents the infection status observation data collected after the β-times propagation process, Indicates that after the lth propagation process ends, node v i Whether infected, the value is 1 for infected, the value is 0 for uninfected, 1≤i,j≤n, 1≤l≤β.

[0077] In step S1 of this embodiment, the calculation formula for obtaining the evaluation index of the parent-child relationship between any two nodes is:

[0078]

[0079] in:

[0080]

[0081] In the formula, PRI(v j →v i ) represents node v j is node v i The possibility of the parent node, and They represent the node v after the lth propagation process ends. i and node v j The infection state value; β represents the number of propagation processes, l represents the lth propagation process; I(·) is an indicator function, which takes the value of 1 when the condition in the brackets is met and takes the value of 0 when it is not met; γ ab are predefined weight parameters, a∈[0,1], b∈[0,1].

[0082] Preferably, in step S1, γ ab ∈[-10,10].

[0083] In step S2 of this embodiment, the value of PRI is larger when there is an influence relationship between nodes, and smaller when there is no influence relationship. Most of the existing work uses the hard clustering method K-means to divide the value of the evaluation index into two categories, and uses the value of the boundary position of the two categories as the threshold for determining whether a directed edge exists. However, this hard clustering method may misjudge the evaluation index near the threshold, and there will be large errors when the distinction between the edges of the two categories is not very obvious, thereby reducing the recall rate or precision of the reasoning result, resulting in the final network structure being inaccurate. Therefore, the present invention uses soft clustering, i.e., fuzzy clustering method, to improve the limitations of hard clustering method when pruning.

[0084] The method of obtaining the propagation topology matrix set based on the parent-child relationship evaluation index between any two nodes combined with the soft clustering method is:

[0085] Step S201: clustering all parent-child relationship evaluation indicators using a fuzzy C-means clustering algorithm with a cluster number of 2 to obtain a membership matrix;

[0086] Step S202: Sampling the membership matrix several times to obtain a set of propagation topology matrices.

[0087] Specifically, in step S201, a fuzzy C-means clustering algorithm with a cluster number of 2 is used to cluster all PRIs (v j →v i ) values ​​are clustered to obtain a membership matrix U, where Represents node v i and node v j The membership degree obtained by clustering the PRI indicators is as follows: Indicates the degree of membership in the category with a smaller PRI index value. indicates the degree of membership in the category with a larger PRI index value, and and Satisfaction between

[0088] Specifically, in step S202, the membership matrix U is sampled M times to obtain a propagation topology matrix set IA = {A 1 ,...,A M}, where A m It means that a propagation topology matrix (1≤m≤M) is obtained after the mth sampling.

[0089] Preferably, in step S202, m∈N + .

[0090] In step S202, the membership matrix U is sampled to obtain the propagation topology matrix A m The propagation process is as follows:

[0091] Step S2021: Initialize an n×n propagation topology matrix A with all values ​​set to 0 m ;

[0092] Step S2022: Traverse A in sequence m For each element in, when traversing to the element in row i and column j (A m ) i,j When (A m ) i,j Randomly assign a random number between 0 and 1;

[0093] Step S2023: Traverse A in sequence m For each element in, when traversing to the element in row i and column j (A m ) i,jWhen (A m ) i,j The value is greater than Then (A m ) i,j The value of is set to 1, otherwise (A m ) i,j The value of is set to 0;

[0094] Step S2024: A m The network propagation topology matrix obtained as a sample is returned.

[0095] In step S3 of this embodiment, the method for constructing a corresponding propagation probability matrix for each propagation topology matrix is ​​specifically as follows:

[0096] For each propagation topology matrix A m ∈IA, construct a corresponding n×n propagation probability matrix W m , and then add all the propagation probability matrices to the propagation probability matrix set IW, that is, IW = {W 1 ,...,W M}(1≤m≤M), where according to A m To construct W m The specific implementation is:

[0097]

[0098] In the formula, (A m ) i,j Indicates A m The value of the element in the i-th row and j-th column; rand(0,1) represents a random number between 0 and 1; (W m ) i,j W m The value of the element in the i-th row and j-th column.

[0099] In step S4 of this embodiment, the core idea of ​​the genetic algorithm is to iteratively evolve individuals in the population by simulating the natural evolutionary propagation process to find the optimal solution to the optimization problem, mainly relying on operations such as mutation and crossover to evolve a series of candidate solutions to the optimal state. This embodiment draws on the idea of ​​the genetic algorithm and performs specific operations on the two propagation processes, so that the propagation probability matrix can be closer to the optimal solution.

[0100] The method of performing genetic algorithm operation in this embodiment to obtain a new propagation probability matrix and a corresponding new propagation topology matrix is:

[0101] Step S401: performing a mutation operation on each propagation probability matrix to obtain a first updated propagation probability matrix;

[0102] Step S402: performing a crossover operation on each first updated propagation probability matrix to obtain a second updated propagation probability matrix;

[0103] Step S403: performing an update operation on each second updated propagation probability matrix to obtain a third updated propagation probability matrix;

[0104] Step S404: using the third updated propagation probability matrix as a new propagation probability matrix;

[0105] Step S405: Obtain a new propagation topology matrix based on the new propagation probability matrix.

[0106] Specifically, in step S401, for each propagation probability matrix W in IW m ∈IW performs mutation operation, and the n×n matrix obtained after mutation operation is recorded as Get the first updated propagation probability matrix The calculation formula is:

[0107]

[0108] In the formula, represents the first updated propagation probability matrix The element in the i-th row and j-th column in ; F is a preset scaling factor; (W m ) r1,r2 Represents the matrix W m The element in the r1th row and r2th column of , and its value satisfies r1≠i,r2≠j; (W m ) r3,r4 Represents the matrix W m The element in the r3th row and r4th column of , and its value satisfies r3≠i,r3≠r1,r4≠j,r4≠r2; (W m ) r5,r6 Represents the matrix W m The element in the r5th row and r6th column of , and its value satisfies r5≠i,r5≠r1,r5≠r3,r6≠j,r6≠r2,r6≠r4;

[0109] In step S402, the first updated propagation probability matrix is ​​obtained after the mutation operation Perform a crossover operation and record the n×n matrix obtained after the crossover operation as Get the second updated propagation probability matrix The calculation formula is:

[0110]

[0111] In the formula, represents the second updated propagation probability matrix The element in the i-th row and j-th column, rand(0,1) represents a random number between 0 and 1; CR is a preset crossover probability parameter;

[0112] In step S403, the propagation probability matrix is ​​updated for each second Update, and record the third updated propagation probability matrix of n×n obtained by the update operation as Get the third updated propagation probability matrix The calculation formula is:

[0113]

[0114]

[0115] in, represents the second updated propagation probability matrix The element in the i-th row and j-th column of , μ is a preset parameter; Indicates that after the lth propagation process ends, node v j The infection status value of m ) i,j Indicates A m The value of the element in row i and column j, A m represents the propagation topology matrix, β represents the number of propagation processes, l represents the lth propagation process, Indicates After the second propagation process, node v t The infection status value of (A m ) t,j Indicates A m The value of the element in row t and column j, represents the second updated propagation probability matrix The element at the tth row and jth column in .

[0116] Preferably, in step S401, F∈(0,1).

[0117] Preferably, in step S402, CR∈(0,1).

[0118] In step S405 of this embodiment, for each propagation topology matrix A in IA m Update, and record the n×n matrix obtained after the update operation as Updated new propagation topology matrix The element in row i and column j of The specific calculation method is:

[0119]

[0120] Where μ is a preset parameter.

[0121] Preferably, in step S402, μ∈(0,10000).

[0122] In step S5 of this embodiment, an iterative update strategy is adopted, and the replacement condition needs to be determined. Therefore, an adaptability function needs to be constructed to determine whether to replace. The calculation formula of the adaptability function is:

[0123]

[0124] Among them, λ is a pre-set parameter; (W m ) j,i W m The element in the jth row and ith column of ; Indicates that after the lth propagation process ends, node v i The infection status value of .

[0125] Preferably, in step S6, λ∈[0,10000).

[0126] Traverse each propagation topology matrix A in IA in turn m , calculate A m And its corresponding propagation probability matrix W m The fitness function value f(A m ,W m ), and then calculate the updated and The fitness function value of Calculate the difference between two fitness function values If the condition dif is met m >0, then replace A in IA m Replace with Change the W in IW m Replace with Otherwise, the A in IA and IW m and W m remain unchanged,

[0127] In this embodiment, it is necessary to give the condition for stopping the iteration of the genetic algorithm operation to prevent infinite iteration. Therefore, step S7 is also included to traverse all the differences. If the maximum value of the difference is less than the given threshold τ, the iteration is stopped. Otherwise, the genetic algorithm operation is returned to obtain a new propagation probability matrix and a corresponding new propagation topology matrix.

[0128] Specifically, step S7 is to traverse each propagation topology matrix A in IA. m and the corresponding propagation probability matrix W m The calculated fitness function value difference dif mIf the maximum value is less than the given threshold τ, then each propagation topology matrix A in IA is traversed in turn. m , calculate A m And the corresponding propagation probability matrix W m The fitness function value f(A m ,W m ), and the fitness function is taken as f(A m ,W m ) The maximum corresponding propagation topology matrix A m And its corresponding propagation probability matrix W m Return, otherwise return to step S4 and continue the genetic algorithm operation to obtain a new propagation probability matrix and a corresponding new propagation topology matrix.

[0129] In order to facilitate the understanding and implementation of the present invention by ordinary technicians in this field, the present embodiment is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described here are only used to illustrate and explain the present embodiment, and are not used to limit the present invention. Five networks are used (as shown in Table 1), among which networks LFR1, LFR2, LFR3, LFR4, and LFR5 are artificial propagation networks generated by the LFR algorithm. The historical propagation result data of the propagation network is obtained by randomly selecting 15% of the nodes as the initial infection points according to the IC model and performing multiple simulation propagations.

[0130] Table 1 Experimental network

[0131] network Number of nodes Network average degree LFR1 200 4 LFR2 400 4 LFR3 600 4 LFR4 800 4 LFR5 1000 4

[0132] Please see Figure 2 , is a result diagram of the corresponding F value of the propagation network reconstructed on the artificial network generated by the LFR algorithm in a specific implementation manner. It can be seen that a propagation network reconstruction method based on historical propagation results provided in this embodiment can accurately reconstruct the propagation network based on historical infection status result data.

[0133] In summary, this embodiment provides a method for reconstructing a propagation network based on historical propagation results, uses the parent-child relationship evaluation index PRI to measure the possibility of an influence relationship between nodes, and uses the fuzzy C-means clustering algorithm combined with sampling to perform preliminary mining of potential influence relationships between nodes, and uses a genetic algorithm to optimize the initial solution obtained, further improving the accuracy of the reconstructed propagation network. Through the method in this article, the influence relationship between nodes in the propagation network can be relatively accurately found by simply using the historical propagation result data of the node's infection status.

[0134] Example 2

[0135] Embodiment 2 provides a propagation network reconstruction device based on historical propagation results, such as Figure 3 As shown, including:

[0136] The propagation network construction module is used to construct the infection status data propagation network and obtain the parent-child relationship evaluation index between any two nodes;

[0137] The propagation topology matrix acquisition module is used to obtain a propagation topology matrix set based on the parent-child relationship evaluation index between any two nodes combined with the soft clustering method;

[0138] A propagation probability matrix acquisition module is used to construct a corresponding propagation probability matrix for each propagation topology matrix;

[0139] An updating module, used for performing genetic algorithm operation to obtain a new propagation probability matrix and a corresponding new propagation topology matrix;

[0140] An adaptability function building module, used to establish an adaptability function, obtain a first adaptability function value of a propagation topology matrix and a corresponding propagation probability matrix, and obtain a second adaptability function value of a new propagation topology matrix and a corresponding new propagation probability matrix;

[0141] A replacement module is used to compare the first fitness function value and the second fitness function value. If the first fitness function value is greater than the second fitness function value, the propagation topology matrix is ​​replaced by a new propagation topology matrix and the propagation probability matrix is ​​replaced by a new propagation probability matrix; otherwise, the propagation topology matrix and the propagation probability matrix remain unchanged.

[0142] Example 3

[0143] Embodiment 3 provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned method for reconstructing a propagation network based on historical propagation results when executing the program.

[0144] Example 4

[0145] Embodiment 4 provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for reconstructing a propagation network based on historical propagation results is implemented.

[0146] The memory in the embodiment of the present invention is used to store various types of data to support the operation of the electronic device. Examples of such data include: any computer program used to operate on the electronic device.

[0147] The method for reconstructing a propagation network based on historical propagation results disclosed in the embodiment of the present invention can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. In the process of implementing the propagation, each step of the method for reconstructing a propagation network based on historical propagation results can be completed by an integrated logic circuit of hardware in the processor or instructions in the form of software. The above-mentioned processor may be a general-purpose processor, a digital signal processor (DSP, Digital Signal Processor), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The processor can implement or execute the methods, steps and logic block diagrams disclosed in the embodiment of the present invention. The general-purpose processor may be a microprocessor or any conventional processor, etc. In combination with the steps of the method disclosed in the embodiment of the present invention, it can be directly embodied as a hardware decoding processor to execute, or it can be executed by a combination of hardware and software modules in the decoding processor. The software module may be located in a storage medium, which is located in a memory. The processor reads the information in the memory and completes the steps of the method for reconstructing a propagation network based on historical propagation results provided in the embodiment of the present invention in combination with its hardware.

[0148] In an exemplary embodiment, the electronic device may be implemented by one or more application specific integrated circuits (ASIC), DSP, programmable logic device (PLD), complex programmable logic device (CPLD), FPGA, general purpose processor, controller, microcontroller (MCU), microprocessor, or other electronic components to execute the aforementioned method.

[0149] It can be understood that the memory can be a volatile memory or a non-volatile memory, and can also include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a magnetic random access memory (FRAM), a flash memory, a magnetic surface memory, an optical disk, or a compact disc read-only memory (CD-ROM); the magnetic surface memory can be a disk memory or a tape memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of RAM are available, such as static random access memory (SRAM), synchronous static random access memory (SSRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), direct memory bus random access memory (DRRAM). The memory described in the embodiments of the present invention is intended to include, but is not limited to, these and any other suitable types of memory.

[0150] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for reconstructing a propagation network based on historical propagation results, characterized in that: The following steps are involved: Construct an infection status data propagation network and obtain the parent-child relationship evaluation index between any two nodes; Based on the parent-child relationship evaluation index between any two nodes and the soft clustering method, a propagation topology matrix set is obtained; Construct a corresponding propagation probability matrix for each propagation topology matrix; Perform genetic algorithm operation to obtain a new propagation probability matrix and a corresponding new propagation topology matrix; Establishing an adaptability function, obtaining a first adaptability function value of a propagation topology matrix and a corresponding propagation probability matrix, and obtaining a second adaptability function value of a new propagation topology matrix and a corresponding new propagation probability matrix; The difference between the first fitness function value and the second fitness function value is obtained. If the difference is greater than 0, the propagation topology matrix is ​​replaced by the new propagation topology matrix and the propagation probability matrix is ​​replaced by the new propagation probability matrix. Otherwise, the propagation topology matrix and the propagation probability matrix remain unchanged.

2. The method for reconstructing a propagation network based on historical propagation results according to claim 1, characterized in that: The calculation formula for the parent-child relationship evaluation index between any two nodes based on the propagation network is: in: 1≤i,j≤n In the formula, PRI(v j →v i ) represents node v j is node v i The possibility of the parent node, n is the number of nodes in the propagation network, and They represent the node v after the e-th propagation process ends. i and node v j The infection state value, β represents the number of propagation processes; I(·) is an indicator function, which takes the value of 1 when the condition in the brackets is met and takes the value of 0 when it is not met; γ ab are predefined weight parameters, a∈[0,1], b∈[0,1].

3. The method for reconstructing a propagation network based on historical propagation results according to claim 1, characterized in that: The method of obtaining the propagation topology matrix set based on the parent-child relationship evaluation index between any two nodes combined with the soft clustering method is: The fuzzy C-means clustering algorithm with a cluster number of 2 is used to cluster all parent-child relationship evaluation indicators to obtain the membership matrix U, where Represents node v i and node v j The membership degree between them is obtained by clustering the parent-child relationship evaluation index. Indicates the degree of membership of the category with a smaller value of the parent-child relationship evaluation index. represents the degree of membership of the category with a larger value of the parent-child relationship evaluation index, and and Satisfaction between 1≤i,j≤n, n is the number of nodes in the propagation network; The membership matrix is ​​sampled several times to obtain a set of propagation topology matrices.

4. The method for reconstructing a propagation network based on historical propagation results according to claim 1, characterized in that: The calculation formula for constructing the corresponding propagation probability matrix for each propagation topology matrix is: In the formula, A m is the propagation topology matrix, W m is the propagation probability matrix, (A m ) i,j Indicates A m The value of the element in the i-th row and j-th column; rand(0,1) represents a random number between 0 and 1; (W m ) i,j W m The value of the element in the i-th row and j-th column, n is the number of nodes in the propagation network.

5. The method for reconstructing a propagation network based on historical propagation results according to claim 1, characterized in that: The method of performing genetic algorithm operation to obtain a new propagation probability matrix and a corresponding new propagation topology matrix is: Perform mutation operation on each propagation probability matrix to obtain the first updated propagation probability matrix, and the calculation formula is: Where W m is the propagation probability matrix, represents the first updated propagation probability matrix The element in the i-th row and j-th column in ; F is a preset scaling factor; (W m ) r1,r2 Represents the matrix W m The element in the r1th row and r2th column of , and its value satisfies r1≠i,r2≠j; (W m ) r3,r4 Represents the matrix W m The element in the r3th row and r4th column of , and its value satisfies r3≠i,r3≠r1,r4≠j,r4≠r2; (W m ) r5,r6 Represents the matrix W m The element in the r5th row and r6th column of , and the value satisfies r5≠i,r5≠r1,r5≠r3,r6≠j,r6≠r2,r6≠r4, where n is the number of nodes in the propagation network; The second updated propagation probability matrix is ​​obtained by performing a crossover operation on each first updated propagation probability matrix, and the calculation formula is: In the formula, represents the second updated propagation probability matrix The element in the i-th row and j-th column; CR is a preset crossover probability parameter; rand(0,1) represents a random number generated between 0 and 1; An update operation is performed on each second updated propagation probability matrix to obtain a third updated propagation probability matrix, and the calculation formula is: In the formula, represents the second updated propagation probability matrix The element in the i-th row and j-th column of , μ is a preset parameter; Indicates that after the e-th propagation process, node v j The infection state value, β represents the number of propagation processes; (A m ) i,j Indicates A m The value of the element in row i and column j, A m represents the propagation topology matrix; Indicates that after the e-th propagation process, node v t The infection status value of (A m ) t,j Indicates A m The value of the element in row t and column j, represents the second updated propagation probability matrix The element at row t and column j in ; Using the third updated propagation probability matrix as a new propagation probability matrix; Based on the new propagation probability matrix, the new propagation topology matrix is ​​obtained. The calculation formula is: Where μ is a preset parameter.

6. The method for reconstructing a propagation network based on historical propagation results according to claim 1, characterized in that: The calculation formula of the fitness function is: In the formula, A m is the propagation topology matrix, W m is the propagation probability matrix, λ is a pre-set parameter; (W m ) j,i W m The element in the jth row and ith column of ; and They represent the node v after the e-th propagation process ends. i and node v j The infection status value of m ) i,j Indicates A m The value of the element in row i and column j, (A m ) j,i Indicates A m The value of the element in the jth row and the ith column, β represents the number of propagation processes, n is the number of nodes in the propagation network, 1≤i,j≤n.

7. The method for reconstructing a propagation network based on historical propagation results according to claim 1, characterized in that: Traverse all the differences. If the maximum value of the difference is less than the given threshold τ, stop the iteration. Otherwise, return to the genetic algorithm operation to obtain a new propagation probability matrix and a corresponding new propagation topology matrix.

8. A propagation network reconstruction device based on historical propagation results, characterized in that: include: The propagation network construction module is used to construct the infection status data propagation network and obtain the parent-child relationship evaluation index between any two nodes; The propagation topology matrix acquisition module is used to obtain a propagation topology matrix set based on the parent-child relationship evaluation index between any two nodes combined with the soft clustering method; A propagation probability matrix acquisition module is used to construct a corresponding propagation probability matrix for each propagation topology matrix; An updating module, used for performing genetic algorithm operation to obtain a new propagation probability matrix and a corresponding new propagation topology matrix; An adaptability function building module, used to establish an adaptability function, obtain a first adaptability function value of a propagation topology matrix and a corresponding propagation probability matrix, and obtain a second adaptability function value of a new propagation topology matrix and a corresponding new propagation probability matrix; A replacement module is used to obtain the difference between the first fitness function value and the second fitness function value. If the difference is greater than 0, the propagation topology matrix is ​​replaced by the new propagation topology matrix and the propagation probability matrix is ​​replaced by the new propagation probability matrix. Otherwise, the propagation topology matrix and the propagation probability matrix remain unchanged.

9. An electronic device, characterized in that: It includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the method for reconstructing a propagation network based on historical propagation results as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for reconstructing a propagation network based on historical propagation results as described in any one of claims 1 to 7 is implemented.

Images