A method, device and storage medium for constructing a rumor propagation model

Abstract

The present invention discloses a method, device and storage medium for constructing a rumor propagation model, constructing a blank model of a social network, initializing the node state of a user node in the blank model of the social network; determining the intimacy between the user node and its neighbor nodes, determining the personal recognition of the user node according to the recognition degree of the user node to the rumor, and determining the initial propagation node V according to the degree centrality of the user node. i , and obtain the initial social network propagation model; judge the node status of each node in the initial social network propagation model, and update the node status in real time according to the judgment result; and after the current propagation round reaches the threshold number of times, output the social network propagation model with updated node status, that is, the rumor propagation model. In the process of rumor propagation, the node status of the user node in the rumor propagation model is updated in real time to obtain a more realistic rumor propagation model, which improves accuracy.

Description

Technical Field

[0001] The present invention relates to the field of information dissemination technology, and in particular to a method, device and storage medium for constructing a rumor propagation model. Background Art

[0002] The popularity of online social networks has greatly changed people's living habits. The number of Internet users is also increasing rapidly, and people are becoming more and more accustomed to expressing their personal opinions on various social platforms. However, while social networks bring convenience to people, rumors are also spreading wildly. These rumors confuse the authenticity of information, manipulate public opinion, and undermine government authority.

[0003] The rumor propagation and diffusion model involves multiple disciplines and is influenced by many factors. The rapid development of social networks in the past two decades has led to many new features in the rumor propagation media, methods, and forms.

[0004] Current research on rumor propagation is usually based on infectious disease models, which have inherent limitations in dealing with individual heterogeneity. There is a clear difference between rumor propagation and disease propagation. The motivation for disease propagation is relatively simple, while rumor propagation is significantly different. Various driving factors, such as personal factors and environmental factors, will affect the rumor propagation process.

[0005] At present, in the research on rumor driving factors, most researchers have only studied one or two driving factors. Faced with the complex network environment, rumor driving factors are diverse, and traditional rumor models are no longer suitable for propagation modeling in the current scenario.

[0006] Current models do not take into account the influence of individual rumor spreaders and various rumor driving factors on the rumor spreading process, and have many limitations. For example, the state changes of individual rumor spreaders during the rumor spreading process cannot be obtained in real time, resulting in low accuracy of the established model. Summary of the invention

[0007] The technical problem to be solved by the present invention is that the current models do not take into account the influence of the spreading individuals and various rumor driving factors on the rumor spreading process. The state changes of the spreading individuals during the rumor spreading process cannot be obtained in real time, and the accuracy of the established model is low. The purpose is to provide a method, device and storage medium for constructing a rumor propagation model. During the rumor spreading process, the node status of user nodes in the social network structure is obtained for real-time update, thereby improving the accuracy of the model and making the obtained rumor propagation model more realistic.

[0008] The present invention is achieved through the following technical solutions:

[0009] The first aspect of the present invention provides a method for constructing a rumor propagation model, comprising the following specific steps:

[0010] S1. construct a blank model of a social network according to user nodes and connection relationships between user nodes, and initialize the node states of user nodes in the blank model of the social network;

[0011] S2. Determine the intimacy between the user node and its neighbor nodes, determine the personal recognition of the user node based on the user node's recognition of rumors, and determine the initial propagation node V based on the degree centrality of the user node. i ;

[0012] The intimacy, the personal recognition and the initial propagation node V i , substitute into the blank model of the social network to obtain the initial social network propagation model;

[0013] S3, judging the node status of each user node of the initial social network propagation model in real time, and updating the node status in real time according to the judgment result;

[0014] If the current propagation round reaches the threshold number of times, output the social network propagation model after the node status is updated, and the social network propagation model after the node status is updated is a rumor propagation model;

[0015] The real-time judgment of the state of each node of the initial social network propagation model includes the following steps:

[0016] S31, according to the initial propagation node V i With neighboring node V j The intimacy degree is obtained to obtain the initial propagation node V i To neighbor node V j The rumor spreading intensity during the rumor spreading process is used to determine V j Whether interested in the rumor, if not interested in the rumor, the current node state does not change, otherwise, execute step S32;

[0017] S32, according to V j The neighbor node set N(V j ) and V j The intimacy level is V j The neighbor node set N(V j ) for V j The influence of public opinion generated, according to V j The neighbor node set N(V j ) for V j The influence of public opinion and V j Personal recognition, get V j Node status.

[0018] The present invention constructs a blank model of a social network through user nodes and the connection relationship between user nodes, initializes the node status of the user nodes in the blank model of the social network; determines the intimacy between the user node and its neighbor nodes, determines the personal recognition of the user node according to the recognition degree of the user node to the rumor, and determines the initial propagation node V according to the degree centrality of the user node. i , and obtain the initial social network propagation model; judge the node status of each user node in the initial social network propagation model, and update the node status in real time according to the judgment result; and after the current propagation round reaches the threshold number of times, output the social network propagation model after the node status is updated, that is, the rumor propagation model; wherein, judging the status of each node in the initial social network propagation model includes the following steps: according to the initial propagation node V i With neighboring node V j The intimacy degree is obtained to obtain the initial propagation node V i To neighbor node V j The rumor spreading intensity during the rumor spreading process is used to determine V j Is the node interested in the rumor? If not, the current node status does not change. Otherwise, according to V j The neighbor node set N(V j ) and V j The intimacy level is V j The neighbor node set N(V j ) for V j The influence of public opinion generated, according to V j The neighbor node set N(V j ) for V j The influence of public opinion and V j Personal recognition, get V j Node status.

[0019] During the rumor propagation process, the node status of the user node in the rumor propagation model is updated in real time to obtain a more realistic rumor propagation model, thereby improving accuracy.

[0020] Furthermore, the node status includes: non-infected, spreader, rumor refutator and recoverer;

[0021] The initialization of the node status of the user nodes in the blank model of the social network is: updating the node status of all user nodes in the blank model of the social network to non-infected persons.

[0022] Furthermore, the initial propagation node V is determined according to the degree centrality of the user node. i The steps include:

[0023] The initial propagation node V iThe number n is set to 1% of the user nodes in the social network propagation model, and the initial propagation node V is calculated. i , traverse all user nodes in the blank model of the social network, and select the 1% nodes with the largest degree centrality as the initial propagation nodes V i , and the initial propagation node V i The node status is updated to propagator.

[0024] Furthermore, the determination V j Are you interested in rumors? Specifically:

[0025] Set neighbor node V j The detection threshold that can detect rumors;

[0026] Combined with the initial propagation node V i To neighbor node V j Rumor propagation intensity and neighbor nodes V in the rumor propagation process j The threshold for detecting rumors is V j Interested in rumors?

[0027] If the rumor propagation intensity is less than the detection threshold, then V j If no rumors are detected, the node status will not change;

[0028] If the rumor spreading intensity is greater than V j The discovery threshold is j The node status is interested in the rumor.

[0029] Further, according to V j The neighbor node set N(V j ) for V j The influence of public opinion and V j Personal recognition, get V j Node status, including:

[0030] Contrast V j The neighbor node set N(V j ) for V j The influence of public opinion and V j Personal identification:

[0031] If V j The influence of public opinion in the environment is greater than his personal recognition, V j Will distinguish rumors, V j The node status of V j Unable to distinguish rumors, V j The node status is updated to propagator.

[0032] Furthermore, the state of each node of the initial social network propagation model is judged in real time, and the following steps are also included:

[0033] S33. When a round of propagation ends, traverse the user nodes whose node states are propagators and rumor refutators after update, and obtain the user node's interest in the received rumor and the recovery threshold at which the user node can be restored to the node state of an uninfected person;

[0034] The user node's interest in the received rumor and the recovery threshold at which the user node can recover to the node state of an uninfected person are combined to determine whether the node state of the user node can be restored to the node state of the uninfected set.

[0035] Furthermore, the step of obtaining the user node's interest in the received rumor specifically includes:

[0036] After traversing the updated node status, the user nodes are spreaders and rumor refutators. If the node V k The node status is the first time to become a spreader or rumor refuter. According to the degree of interest of the crowd in the rumor, the node V is determined. k The interest level in rumors, if the current node V k The node status is not the first time to become a disseminator or rumor refuter. k Assign interest level;

[0037] Update Node V k Interest level, using node V k The updated interest level is compared with the recovery threshold at which the node state of the user node can be restored to the node state of the uninfected set:

[0038] If the node V k If the interest level is less than the recovery threshold, the node status of the user node is updated to a communicator;

[0039] If the node V k If the interest degree is greater than the recovery threshold, the node status remains unchanged.

[0040] Furthermore, the update node V k The steps for determining the degree of interest include:

[0041] When a propagation round is over, the node V whose node status is the propagator is traversed. p , update V p Node status, calculate V p The neighbor node set N(V p ) for V p The impact of public opinion generated:

[0042] If Vp The influence of public opinion in the environment is greater than his personal recognition, V p Will distinguish rumors, will V p The node status of V p The status remains unchanged;

[0043] After traversing and updating, the node status is the user node of the spreader and rumor refuter. For each traversed node V k , update the current node’s interest in the rumor.

[0044] A second aspect of the present invention 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 is used to implement a method for constructing a rumor propagation model when executing the program.

[0045] A third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is used to implement a method for constructing a rumor propagation model.

[0046] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0047] The present invention obtains a more realistic rumor propagation model by updating the status of user nodes in the social network structure in real time during the rumor propagation process, thereby improving accuracy and better describing the process of rumor propagation in the social network. The model can more effectively control rumors and curb the spread of rumors as early as possible. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without creative work. In the drawings:

[0049] Figure 1 A schematic diagram of a method step flow chart in an embodiment of the present invention;

[0050] Figure 2 A social network propagation model including 15 user nodes and 28 edges in an example embodiment of the present invention;

[0051] Figure 3 This is a schematic diagram of the rumor propagation state in a social network at the initial stage in an example embodiment of the present invention;

[0052] Figure 4It is a schematic diagram of rumor propagation status in a running social network in an example embodiment of the present invention;

[0053] Figure 5 It is a schematic diagram of rumor propagation status in a running social network in an example embodiment of the present invention;

[0054] Figure 6 It is a schematic diagram of rumor propagation status in a running social network in an example embodiment of the present invention;

[0055] Figure 7 It is a schematic diagram of rumor propagation status in a running social network in an example embodiment of the present invention;

[0056] Figure 8 It is a schematic diagram of rumor propagation status in a running social network in an example embodiment of the present invention;

[0057] Fig. 9 It is a schematic diagram of the final state of rumor propagation in a social network in an example embodiment of the present invention;

[0058] Fig.10 Schematic diagram of the dynamic process of rumor propagation in a social network in an example embodiment of the present invention. DETAILED DESCRIPTION

[0059] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0060] Example 1

[0061] like Figure 1 As shown, this embodiment provides a method for constructing a rumor propagation model, including the following specific steps:

[0062] S1. construct a blank model of a social network according to user nodes and connection relationships between user nodes, and initialize the node states of user nodes in the blank model of the social network;

[0063] like Figure 2 As shown, the constructed social network blank model is G(V,E), where: V represents the user node set in the social network, and E represents the edge set between nodes in the social network;

[0064] The node states include: non-infected, spreader, rumor refutator and restorer; the node state of the user node in the blank model of the social network is initialized as follows: the node state of all user nodes in the blank model of the social network is updated to non-infected.

[0065] S2. Determine the intimacy between the user node and its neighbor nodes, determine the personal recognition of the user node based on the user node's recognition of rumors, and determine the initial propagation node V based on the degree centrality of the user node. i , and obtain the initial social network propagation model, such as Figure 3 As shown;

[0066] Determining the intimacy between a user node and its neighbor nodes includes the following specific steps:

[0067] Get the Jaccard coefficient between neighboring nodes:

[0068]

[0069] Among them, V i and V j is a pair of neighbor nodes, N(V i ) indicates user V i The set of neighbor nodes of

[0070] Calculate user V using the 0-1 normalization method i V j Degree of intimacy:

[0071]

[0072] Among them, min is the minimum Jaccard coefficient between node Vi and its neighbors, and max is the maximum Jaccard coefficient between node Vi and its neighbors.

[0073] Determining the personal recognition of the user node according to the degree of recognition of the rumor by the user node specifically includes the following steps:

[0074] Distinguished vi ~N(α,β) 2

[0075] Among them, the user's personal identification Distinguish vi Comply with parameters α, β 2 The normal distribution of , α is the mean of the personal recognition of all users, β is the variance of the personal recognition of all users, where the mean α is set to 0.5 and the variance β is set to 0.25.

[0076] Determine the initial propagation node V according to the degree centrality of the user node i , and obtaining the initial social network propagation model specifically includes the following steps:

[0077] The initial propagation node V i The number n is set to 1% of the user nodes in the social network propagation model, and the initial propagation node V is calculated. i, traverse all user nodes in the blank model of the social network, and select the 1% nodes with the largest degree centrality as the initial propagation nodes V i , and the initial propagation node V i Add to the communicator set, the calculation formula is:

[0078]

[0079] in, is node v i The number of neighbors, V represents the user node set in the social network, |V| is the number of nodes in the network; traverse all nodes, select the 1% nodes with the largest degree centrality as the initial propagation nodes, add I0, and select at least one when selecting the initial propagation node.

[0080] S3, judging the node status of each user node in the initial social network propagation model, and updating the node status in real time according to the judgment result;

[0081] If the current propagation round reaches the threshold number of times, output the social network propagation model after the node status is updated, and the social network propagation model after the node status is updated is a rumor propagation model;

[0082] The state of each node in the initial social network propagation model is judged, including the following steps:

[0083] S31, according to the initial propagation node V i With neighboring node V j The intimacy degree is obtained to obtain the initial propagation node V i To neighbor node V j The rumor propagation intensity during the rumor propagation process, combined with the initial propagation node V i To neighbor node V j Rumor propagation intensity and neighbor nodes V in the rumor propagation process j The threshold for detecting rumors is V j Interested in the rumor: If not interested in the rumor, the current node state does not change, otherwise, execute step S32;

[0084] The current propagation round is t, traversing the propagator set I t Gather with rumor refutators C t , for each traversed node V i , if the node enters the set for the first time, calculate V i Current level of interest in the rumor

[0085]

[0086] Among them, the degree of interest of the crowd in rumors The parameters are μ,σ 2 Normal distribution, μ is the mean of the interest level of the crowd, the value of μ is generally 0.5, σ is the variance of the interest level of the crowd, the value of σ is generally 0.25;

[0087] Traverse V i The uninfected neighbor nodes, calculate V i Each node V in the set of uninfected neighbor nodes j , calculated to V i To V j The intensity of rumor propagation

[0088]

[0089] Among them, Importance is the importance of the rumor, θ is the degree of platform involvement, Yes V i With V j the degree of intimacy;

[0090] Set neighbor node V j The detection threshold that can detect rumors;

[0091] Combined with the initial propagation node V i To neighbor node V j Rumor propagation intensity and neighbor nodes V in the rumor propagation process j The threshold for detecting rumors is V j Interested in rumors?

[0092] If the rumor propagation intensity is less than the detection threshold, then V j If no rumors are detected, the node status will not change;

[0093] If the rumor spreading intensity is greater than V j The discovery threshold is j The node status is interested in the rumor.

[0094] S32, according to V j The neighbor node set N(V j ) and V j The intimacy level is V j The neighbor node set N(V j ) for V j The influence of public opinion generated, according to V j The neighbor node set N(V j ) for V j The influence of public opinion and V j Personal recognition, get Vj Node status;

[0095] For each V j , calculate V j The neighbor node set N(V j ) for V j The influence of public opinion

[0096] in, Yes V j All neighbors in the group of spreaders or rumor refutators are V j The influence of public opinion caused by j The node that is the intersection of the neighbor set of the propagator or rumor refutator set, n is the node in V j The node that is the intersection of the neighbor set and the rumor-refuting set. is the neighbor node V m The personal identification of e is the natural logarithm of the mathematical constant. Yes V m With V j the degree of intimacy;

[0097] Contrast V j The neighbor node set N(V j ) for V j The influence of public opinion and V j Personal identification:

[0098] If V j The influence of public opinion in the environment is greater than his personal recognition, V j Will distinguish rumors, V j The node status of V j Unable to distinguish rumors, V j The node status is updated to propagator.

[0099] S33. When a round of propagation ends, traverse the user nodes whose node states are propagators and rumor refutators after update, and obtain the user node's interest in the received rumor and the recovery threshold at which the user node can be restored to the node state of an uninfected person;

[0100] The user node's interest in the received rumor and the recovery threshold at which the user node can recover to the node state of an uninfected person are combined to determine whether the node state of the user node can be restored to the node state of the uninfected set.

[0101] After traversing the updated node status, the user nodes are spreaders and rumor refutators. If the node V kThe node status is the first time to become a spreader or rumor refuter. According to the degree of interest of the crowd in the rumor, the node V is determined. k The interest level in rumors, if the current node V k The node status is not the first time to become a disseminator or rumor refuter. k Assign interest level;

[0102] After traversing the updated node status, the user nodes are spreaders and rumor refutators. If the node V k The node status is the first time to become a spreader or rumor refuter. According to the degree of interest of the crowd in the rumor, the node V is determined. k The interest level in rumors, if the current node V k The node status is not the first time to become a disseminator or rumor refuter. k Assign interest level;

[0103] Update Node V k The updating step specifically includes:

[0104]

[0105] Among them, r is the decay degree of user interest; using node V i The updated interest level is compared with the set recovery threshold Recover, and the value of Recover is generally 0.3;

[0106] Using Node V k The updated interest level is compared with the recovery threshold at which the node state of the user node can be restored to the node state of the uninfected set:

[0107] If the node V k If the interest level is less than the recovery threshold, the node status of the user node is updated to a communicator;

[0108] If the node V k If the interest degree is greater than the recovery threshold, the node status remains unchanged.

[0109] When a propagation round is over, the node V whose node status is the propagator is traversed. p , update V p Node status, calculate V p The neighbor node set N(V p ) for V p The impact of public opinion generated:

[0110] If V p The influence of public opinion in the environment is greater than his personal recognition, V p Will distinguish rumors, V pThe node status of V p The status remains unchanged;

[0111] After traversing and updating, the node status is the user node of the spreader and rumor refuter. For each traversed node V k , update the current node’s interest in the rumor.

[0112] Among them, a propagation round is: the initial propagation node V i To the initial propagation node V i Neighbor node V j After spreading the rumor, the neighbor node V j The node status is updated, and the neighbor node V j Before becoming a new rumor spreading node or a new rumor refuting node, Figures 2 to 4 , is a propagation round.

[0113] According to the number of different user set nodes in each stage, a propagation curve is drawn; the propagation process of rumors in the social network is restored through the curve and the network status data at each moment, so as to more effectively control and limit the spread of rumors.

[0114] In some possible embodiments, the recovery threshold is set to 0.3, the interest decay value is set to 0.8, the discovery threshold is set to 0.5, the rumor importance is set to 1, the platform intervention level is set to 1, the mean μ of the rumor interest level is 0.5, and the variance σ is 0.25.

[0115] In some possible embodiments, the present embodiment also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor is used to implement a method for constructing a rumor propagation model when executing the program.

[0116] In some possible embodiments, the present embodiment further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is used to implement a method for constructing a rumor propagation model.

[0117] Example 2

[0118] This embodiment provides a specific construction process of a rumor propagation model that takes into account individual heterogeneity and multi-factor driving:

[0119] Input network structure data and traverse all nodes, at this time V=0;

[0120] Calculate the degree centrality of node V, calculate the closeness between node V and its neighbors, and calculate the personal recognition of node v;

[0121] Determine whether V is completely traversed:

[0122] If V is not completely traversed, set V = V + 1 and jump to step: calculate the degree centrality of node V;

[0123] If V is completely traversed, initialize the number of propagation rounds T = 0;

[0124] Traverse the user nodes in the set of disseminators and rumor refutators in round T, at this time I=0

[0125] Determine whether I enters the collection for the first time:

[0126] If I is not entering the set for the first time, assign the interest level of node I;

[0127] If I enters the set for the first time, set the interest level of node I;

[0128] Traverse all uninfected neighbor nodes J=0 of the current propagation node I;

[0129] Determine whether the rumor propagation intensity around the current node J is greater than the discovery threshold;

[0130] If the rumor propagation intensity around the current node J is greater than the discovery threshold, the state of node J remains unchanged, J=J+1, until the rumor propagation intensity around the current node J is greater than the discovery threshold;

[0131] Determine whether the sum of the public opinion influence around the disseminated node and the personal recognition is greater than zero:

[0132] If the sum of the influence of public opinion around the node being spread and the personal recognition is less than zero, the node being spread has not identified the rumor, and J is added to the set of spreaders;

[0133] If the sum of the influence of public opinion around the node being spread and the personal recognition is greater than zero, the node being spread will identify the rumor and add J to the set of rumor debunkers;

[0134] Determine whether J is completely traversed;

[0135] Update node interest level;

[0136] Determine whether the interest level of node I is greater than the recovery threshold;

[0137] If the interest level of node I is less than the recovery threshold, determine whether I is completely traversed;

[0138] If the interest level of node I is greater than the recovery threshold, I is added to the recovery state set to determine whether I has been traversed completely:

[0139] If I is not completely traversed, then I = I + 1, jump to step: determine whether I is the first set;

[0140] If I is completely traversed, all propagator nodes are traversed, M = 0;

[0141] Determine whether the sum of the public opinion influence and personal recognition around the disseminated node is greater than zero:

[0142] If the sum of the influence of public opinion and personal recognition around the node being spread is greater than zero, the node identifies the rumor and joins the rumor-refuting node;

[0143] If the sum of the influence of public opinion and personal recognition around the node being spread is less than zero, then the node has not identified the rumor and remains in its original state;

[0144] Determine whether M is completely traversed:

[0145] If M is not completely traversed, then M=M+1, and jump to step: determine whether the sum of the public opinion influence and personal recognition around the propagated node is greater than zero;

[0146] If M is completely traversed, the update propagation round is T = T + 1;

[0147] Determine whether T is greater than the set value K or whether there are new neighbors and infected people;

[0148] If T is less than the set value K or there are new neighbors and infected people, then T = T + 1, jump to step: traverse each node in the set of spreaders and rumor refutators in the Tth round, I = 0;

[0149] If T is greater than the set value K or whether there are new neighbors and infected people, the rumor propagation process is output.

[0150] Example 3

[0151] This embodiment provides a specific implementation example based on Embodiment 1 and Embodiment 2:

[0152] 1. Construct a blank model of social network as G(V,E), and the user nodes in the social network are V = {v1,v2,…,v 15}, the edge between user nodes is E = {(v1,v2),(v1,v3),…,(v 12 ,v 15 )},

[0153] Initialize the node status of the user node, which includes: non-infected, spreader, rumor refuter and restorer; initialize the node status of the user node in the blank model of the social network: update the node status of all user nodes in the blank model of the social network to non-infected. Construct the user node set according to the user's node status: non-infected set I0, spreader set S0, rumor refuter set C0 and restorer set R0.

[0154] For each node V i, calculate V between each pair of neighbors according to the social network blank model i , V j The degree of intimacy is calculated to calculate the personal recognition of the user node for the degree of identification of rumors Given the degree of importance of the rumor in the rumor propagation process as 1, the degree of platform intervention in the rumor propagation process as 1, the detection threshold of the user node to detect the rumor as 0.4, the recovery threshold of the user node to return to the state of the propagator node after becoming a disseminator or rumor debunker as 0.3, and the attenuation degree of the user node's interest in the rumor as 0.8.

[0155] The important information of each node is shown in Table 1 (some node data needs to be generated during the operation, here all information is summarized after the operation, some nodes do not participate in the transmission throughout the process, and their data are replaced by -).

[0156] Table 1 User node information

[0157]

[0158]

[0159] 2. If Figure 2 As shown, follow step 2 to calculate the degree centrality of all nodes, select the 1% nodes with the largest degree centrality, and V1 has the largest degree centrality in the network, which is 0.43. Put node V1 from S0 to the communicator set I0. The node status is as follows Figure 3 shown.

[0160] 3. According to steps S31 and S32, first traverse all nodes {V1} contained in S0 and I0. V1 is the only user node in S0. V1 enters the set for the first time. Calculate the interest level of V1 compared with the rumor. The interest level of node V1 is greater than the recovery threshold. Therefore, traverse all non-infected neighbor nodes of V1 {V2, V3, V4, V6, V7, V8, V9, V10, V11, V12, V13, V14, V15, V16, V17, V18, V19, V20, V21, V22, V23, V24, V25, V26, V27, V28, V29, V30, V31, V32, V33, V34, V45, V66, V19, V21, V22, V23, V34, V45, V66 10 ,V 11}, calculate the distance V1 sends to each of its uninfected neighbor nodes V j Intensity of the rumor spread Vi,Vj The information of each node is shown in Table 2:

[0161] Table 2: Status of affected nodes during the first round of propagation

[0162] node v1's intimacy with the node The intensity of the transmission Personal identification Influence of public opinion Interest level <![CDATA[V2]]> 1 1 0.58 -0.60 - <![CDATA[V3]]> 0 - 0.43 - - <![CDATA[V4]]> 0.66 0.66 0.55 -0.48 - <![CDATA[V6]]> 0.13 0.13 0.49 - - <![CDATA[V 10 ]]> 0.66 1 0.38 -0.60 - <![CDATA[V 11 ]]> 0.81 0.81 0.51 -0.52 -

[0163] Where {V2,V4,V 10 ,V 11}The rumor propagation intensity received by the node is greater than the discovery threshold (usually 0.4), indicating that the node perceives the rumor and enters the rumor identification process, calculating the public opinion influence of the node's neighbor nodes on it. {V4} If the sum of the node's public opinion influence and the node's personal recognition is greater than 0, it means that the node has identified the rumor and put the node into the rumor debunker set C1; {V2,V 10 ,V 11 If the sum of the node's public opinion influence and the node's personal recognition is less than 0, it means that the node has not identified the rumor and puts the node into the disseminator set S1. The node status is as follows Figure 4 As shown.

[0164] After the first round of propagation ends, the second round of propagation begins, traversing C1 and S1, assigning initial values ​​to each node without interest level; updating the interest level of nodes that already have interest level, among which the interest level of V1 is updated to 0.62;

[0165] Enter the second round and traverse all nodes {V1,V2,V4,V 10 ,V 11}, calculate and update its interest level, traverse the non-infected neighbors of the node, and the node set is {V3,V5,V6,V8,V 13 ,V 14}, calculate the maximum propagation intensity and public opinion influence of the neighbor node set, as shown in Table 3:

[0166] Table 3. Affected node status during the second round of propagation

[0167] node Maximum transmitted intensity Personal identification Influence of public opinion Interest level <![CDATA[V1]]> - 0.40 - 0.62 <![CDATA[V2]]> - 0.58 -0.60 0.54 <![CDATA[V3]]> 0 0.43 - - <![CDATA[V4]]> - 0.55 -0.48 0.40 <![CDATA[V5]]> 0.18 0.62 - - <![CDATA[V6]]> 0.40 0.49 -0.86 - <![CDATA[V8]]> 0.80 0.39 0.47 - <![CDATA[V 10 ]]> - 0.38 -0.60 0.57 <![CDATA[V 11 ]]> - 0.51 -0.52 0.41 <![CDATA[V 13 ]]> 0.42 0.44 -0.18 - <![CDATA[V 14 ]]> 0 0.53 - -

[0168] In the second round of propagation, {V6, V8, V 13 The maximum dissemination intensity of the three nodes is greater than the detection threshold, and enters the rumor judgment state. By calculating the sum of public opinion influence and personal recognition, {V6} is added to the disseminator set I2, {V8, V 13} is added to the rumor-refuting set C2. After the neighbor node judgment is completed, the results are as follows: Next, the status of all infected nodes is re-judged and their surrounding influence is calculated, where {V1,V 10} Successfully affected by the environment, it changes to the rumor-refuting state. The state at the end of the second round of dissemination is as follows Figure 6 shown.

[0169] At the beginning of the third round, the interest level of each node is calculated and updated, {V4,V 11}The interest level is less than the recovery threshold, exit the dissemination, and join the restorer set. The maximum dissemination intensity of {V9} is greater than the discovery threshold, and enters the rumor judgment state. The influence of public opinion around it is calculated. If it is greater than 0, it joins the rumor refutation set. All disseminators are conditionally judged and their surrounding influence is calculated. Among them, {V2} is successfully affected by the environmental state and changes to the rumor refutation state, such as Figure 6 shown.

[0170] The fourth round begins 13}The interest level is less than the recovery threshold, exit the propagation, and join the restorer set. There is no new neighbor node. The status is as follows Figure 7 shown.

[0171] The fifth round begins, {V2, V6, V9, V 14}The interest level is less than the recovery threshold, exit the propagation, and join the restorer set. There is no new neighbor node, the state is as follows Figure 8 shown.

[0172] The sixth round begins, {V1, V8, V 10}The interest level is less than the recovery threshold, exit the propagation, and join the restorer set. There is no new neighbor node, the state is as follows Fig. 9 shown.

[0173] The current communicator set is empty, the judgment ends, a process diagram is drawn according to the node propagation process, and the process ends.

[0174] like Fig.10 The figure shows the process of rumor propagation in social networks.

[0175] Figure 3-Figure 9 In Figure 6 For example, the nodes with the same color as nodes 1, 8, 10, 13, and 14 are propagator nodes, the nodes with the same color as nodes 4 and 11 are restorer nodes, the nodes with the same color as nodes 2 and 6 are rumor-refuting nodes, and the other nodes are uninfected nodes.

[0176] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for constructing a rumor propagation model, characterized in that: The specific steps include: S1. construct a blank model of a social network according to user nodes and connection relationships between user nodes, and initialize the node states of user nodes in the blank model of the social network; S2. Determine the intimacy between the user node and its neighbor nodes, determine the personal recognition of the user node based on the user node's recognition of rumors, and determine the initial propagation node V based on the degree centrality of the user node. i ; The intimacy, the personal recognition and the initial propagation node V i , substitute into the blank model of the social network to obtain the initial social network propagation model; S3, judging the node status of each user node of the initial social network propagation model in real time, and updating the node status in real time according to the judgment result; If the current propagation round reaches the threshold number of times, output the social network propagation model after the node status is updated, and the social network propagation model after the node status is updated is a rumor propagation model; The real-time judgment of the state of each node of the initial social network propagation model includes the following steps: S31, according to the initial propagation node V i With neighboring node V j The intimacy degree is obtained to obtain the initial propagation node V i To neighbor node V j The rumor spreading intensity during the rumor spreading process is used to determine V j Whether interested in the rumor, if not interested in the rumor, the current node state does not change, otherwise, execute step S32; S32, according to V j The neighbor node set N(V j ) and V j The intimacy level is V j The neighbor node set N(V j ) for V j The influence of public opinion generated, according to V j The neighbor node set N(V j ) for V j The influence of public opinion and V j Personal recognition, get V j Node status.

2. The method for constructing a rumor propagation model according to claim 1, characterized in that: The node status includes: non-infected, spreader, rumor refutator and recoverer; The initialization of the node status of the user nodes in the blank model of the social network is: updating the node status of all user nodes in the blank model of the social network to non-infected persons.

3. The method for constructing a rumor propagation model according to claim 2, characterized in that: The initial propagation node V is determined according to the degree centrality of the user node i The steps include: The initial propagation node V i The number n is set to 1% of the user nodes in the social network propagation model, and the initial propagation node V is calculated. i , traverse all user nodes in the blank model of the social network, and select the 1% nodes with the largest degree centrality as the initial propagation nodes V i , and the initial propagation node V i The node status is updated to propagator.

4. The method for constructing a rumor propagation model according to claim 3, characterized in that: The judgment V j Are you interested in rumors? Specifically: Set neighbor node V j The detection threshold that can detect rumors; Combined with the initial propagation node V i To neighboring node V j Rumor propagation intensity and neighbor nodes V in the rumor propagation process j The threshold for detecting rumors is V j Interested in rumors? If the rumor propagation intensity is less than the detection threshold, then V j If no rumors are detected, the node status will not change; If the rumor spreading intensity is greater than V j The discovery threshold is V j The node status is interested in the rumor.

5. The method for constructing a rumor propagation model according to claim 3, characterized in that: According to V j The neighbor node set N(V j ) for V j The influence of public opinion and V j Personal recognition, get V j Node status, including: Contrast V j The neighbor node set N(V j ) for V j The influence of public opinion and V j Personal identification: If V j The influence of public opinion in the environment is greater than his personal recognition, V j Will distinguish rumors, V j The node status of V j Unable to distinguish rumors, V j The node status is updated to propagator.

6. The method for constructing a rumor propagation model according to claim 3, characterized in that: The state of each node of the initial social network propagation model is judged in real time, and the following steps are also included: S33. When a round of propagation ends, traverse the user nodes whose node states are propagators and rumor refutators after update, and obtain the user node's interest in the received rumor and the recovery threshold at which the user node can be restored to the node state of an uninfected person; The degree of interest of the user node in the received rumor and the recovery threshold at which the user node can recover to the node state of an uninfected person are combined to determine whether the node state of the user node can be recovered to the node state of an uninfected person.

7. The method for constructing a rumor propagation model according to claim 6, characterized in that: The step of obtaining the user node's interest level in the received rumor specifically includes: After traversing the updated node status, the user nodes are spreaders and rumor refutators. If the node V k The node status is the first time to become a spreader or rumor refuter. According to the degree of interest of the crowd in the rumor, the node V is determined. k The interest level in rumors, if the current node V k The node status is not the first time to become a disseminator or rumor refuter. k Assign interest level; Update Node V k Interest level, using node V k The updated interest level is compared with the recovery threshold at which the node state of the user's node can be restored to the node state of an uninfected person: If the node V k If the interest level is less than the recovery threshold, the node status of the user node is updated to a communicator; If the node V k If the interest degree is greater than the recovery threshold, the node status remains unchanged.

8. The method for constructing a rumor propagation model according to claim 7, characterized in that: The update node V k The steps for determining the degree of interest include: When a propagation round is over, the node V whose node status is the propagator is traversed. p , update V p Node status, calculate V p The neighbor node set N(V p ) for V p The impact of public opinion generated: If V p The influence of public opinion in the environment is greater than his personal recognition, V p Will distinguish rumors, V p The node status of V p The status remains unchanged; After traversing and updating, the node status is the user node of the spreader and rumor refuter. For each traversed node V k , update the current node’s interest in the rumor.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, it implements a method for constructing a rumor propagation model as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, a method for constructing a rumor propagation model as described in any one of claims 1 to 8 is implemented.

Images