Modeling method and device for non-real public opinion information propagation kinetic model

Abstract

The invention discloses a modeling method and device for a non-real public opinion information propagation dynamic model, belongs to the technical field of non-real public opinion information propagation, and particularly relates to modeling of a propagation dynamic model of online social network information. The problem that the propagation of the online social network information cannot be accurately modeled in the existing research on the propagation and management and control of the online social network information is solved; the method comprises the following steps of: constructing a directed graph network of an online social network according to an attention mechanism of the online social network; and a step for constructing a propagation dynamical model. The modeling method and device of the non-real public opinion information propagation dynamic model are suitable for constructing the non-real public opinion information propagation dynamic model of the heterogeneous directed online social network.

Description

Technical Field

[0001] The present invention relates to the technical field of false public opinion information dissemination, and in particular to the modeling of a dissemination dynamics model of online social network information. Background Art

[0002] The popularity of smart mobile devices and the rapid development of Internet technology have made online social networks an indispensable means of communication for people to receive news and real-time information. Online social networks have the characteristics of open registration, freedom of information release, and heterogeneity of users. Online social network users can receive and send news, consultations, and popular science in any network in real time. However, not all online social network information is verified, which makes users vulnerable to false public opinion information when facing public emergencies.

[0003] The online social network space is anonymous, interactive and communicative, which makes it easier for users to obtain information in this network environment, but at the same time it is difficult to guarantee the reliability and authenticity of the content in the online social network. The emergence of this phenomenon has brought huge challenges to the information security of online social networks and the maintenance of a clear network space.

[0004] Specifically, the uncertainty and anonymity of information in online social networks make the virtual crowd in cyberspace unconstrained. In the face of public emergencies, some false public opinion information may appear in online social networks. The rapid spread of false public opinion information in social networks can easily produce huge negative effects.

[0005] It can be seen that the understanding and control of the laws of the dissemination of false public opinion information has become a major problem that needs to be solved urgently in this field. Exploring the modeling and control strategies of the dissemination of false public opinion information on online social networks has very important theoretical value and practical application value.

[0006] Compared with the slow updating characteristics of traditional paper media and television media, online social networks have the ability to quickly iterate and evolve information at a high speed. When a public opinion event occurs, online social network users will quickly gather various information on the network, and thus form a special model of online social network public opinion information dissemination with the public opinion event as the core, various online social network platforms as carriers, and the opinions, emotions and attitudes of network platform users as the expression form, further generating a special model of online social network public opinion information dissemination that includes interactivity, anonymity and immediacy.

[0007] Under this special mode, the spread of false public opinion information on online social networks is characterized by rapid evolution, wide dissemination range, and profound negative impact. As a result, traditional communication research has seriously lagged behind the dissemination mode of false public opinion information under the influence of online social networks. This has put forward new requirements on how to restore this false public opinion information dissemination mode and design corresponding management and control strategies.

[0008] Although previous studies have conducted in-depth discussions on the dissemination and control of online social network information, most of the research has focused on the qualitative level, and the means of control of online social networks are mostly based on group control strategies, lacking more efficient and targeted control solutions. Therefore, how to accurately model the dissemination of online social network information, how to accurately control false public opinion information on online social networks, and thus improve the efficiency of public opinion control are important challenges facing public opinion control. Summary of the invention

[0009] The present invention proposes a modeling method and device for a false public opinion information propagation dynamics model, which solves the problem that existing research on the propagation and control of online social network information cannot accurately model the propagation of online social network information.

[0010] The modeling method of the false public opinion information propagation dynamics model described in the present invention has the following technical solutions:

[0011] The method comprises the following steps:

[0012] A step for constructing a directed graph network of an online social network according to a follow mechanism of the online social network;

[0013] The directed graph network includes multiple nodes and multiple edges connecting any two account nodes; the nodes represent independent users in the directed online social network; the edges represent the attention relationship between subscribers and subscribers in the directed online social network; the directed graph network is in, represents the set of nodes; ε represents the set of edges, and

[0014] The directed graph network satisfies the following assumptions:

[0015] Hypothesis 1: Based on the Barabási-Albert model, the topological structure of the directed graph network remains approximately stable during the propagation of false public opinion information in social networks;

[0016] Assumption 2: There are no self-loops or repeated connections between any account nodes in the directed graph network;

[0017] Steps for constructing a propagation dynamics model: The steps for constructing a propagation dynamics model include the following steps:

[0018] Steps for defining propagation rules for the spread of false public opinion information;

[0019] Steps for obtaining the relationship between the probability of a node user unknown becoming a propagator and edge characteristics in an online social network:

[0020] Calculate the influence of neighbor nodes on the current node and the steps of the dynamic equations for the propagation of false public opinion information in online social networks.

[0021] Further, a preferred embodiment is provided, wherein the step of constructing a directed graph network of an online social network according to the attention mechanism of the online social network comprises the following steps:

[0022] Steps for node classification:

[0023] The account nodes in the directed graph network are divided into four types of nodes: unknown, hesitant, propagator and immune;

[0024] The unknown persons are online social network users who have not been exposed to false public opinion information so far and are still unaware of the public opinion topic;

[0025] The hesitant ones are online social network users who have learned about false public opinion information but have not yet decided whether to express their opinions or views;

[0026] The disseminator is an online social network user who spreads false public opinion information, opinions or views, represented by I;

[0027] The immune people are online social network users who know the false public opinion information but eventually lose interest in it and stop spreading it;

[0028] Steps to simplify a directed graph network:

[0029] Let S represent the proportion of unknown people in the whole group, let H represent the proportion of hesitant people in the whole group, let I represent the proportion of spreaders in the whole group, let R represent the proportion of immune people in the whole group, and satisfy the following formula:

[0030]

[0031] At time t=0, the group size of communicators who initiate discussion and dissemination of false public opinion information in the social network is assumed to be ignored:

[0032]

[0033] The initial value of the node state ratio of the directed online social network public opinion information dissemination system is approximately:

[0034] S(0)=1, H(0))=I(0)=R(0)=0;

[0035] The step used for the degree distribution.

[0036] Further, a preferred embodiment is provided, wherein the step for degree distribution comprises:

[0037] The degree of a node in a network refers to the number of edges connected to this node;

[0038] In a directed graph network, the in-degree of a node refers to the number of edges entering the node, which indicates the number of users followed by other users; the out-degree of a node refers to the number of edges departing from the node, which indicates the number of users followed by the user;

[0039] For any class A∈{S,H,I,M,R}:

[0040] Let A jk (t) represents the proportion of online social network users in state A with in-degree j∈{1, ..., J} and out-degree k∈{1, ..., K} at time t;

[0041] Let p jk The probability that an arbitrary node has degree (j, k) is expressed as follows:

[0042]

[0043] p jk =S jk (t)+H jk (t)+I jk (t)+R jk (t);

[0044] The probability of randomly selecting a node with a certain in-degree and out-degree is described by the joint probability distribution F(x, y) of the degrees and expressed as a probability generating function:

[0045]

[0046] Where x and y are placeholders for constructing the probability generating function;

[0047] Average in-degree of directed graph networks <j>and average out-degree <k>It is expressed by the following formulas and symbols:

[0048]

[0049]

[0050]

[0051] Where m is the total number of in-degrees in the online social network, which is the same as the total number of out-degrees; y is the total number of nodes in the online social network.

[0052] Furthermore, a preferred implementation is provided, wherein the step of defining a propagation rule for propagating false public opinion information comprises:

[0053] The propagation rules are as follows:

[0054] Rule 1: Let an S jk Type Node Online social network users receive false public opinion information from one of their disseminator neighbors, which is a Poisson process with a constant rate r;

[0055] When the S jk When a user at the type node is exposed to a false public opinion topic, the user's state will immediately change to a hesitant state;

[0056] Rule 2: H jk Type node users need to decide whether to spread false public opinion information;

[0057] Among them, the probability of choosing to spread information related to false public opinion topics is γ, and the user's status will be changed to a disseminator;

[0058] The probability of remaining silent and ignoring the public opinion information is 1-γ, and the user's status will be changed to immune;

[0059] Rule 3: One I jk Type node users may completely lose interest in the public opinion information and become immune at a rate of μ.

[0060] Further, a preferred embodiment is provided, wherein the step of obtaining the relationship between the probability of an unknown node user becoming a communicator and the edge characteristics in an online social network comprises:

[0061] Let θ(t) represent the probability that the directed edge with the previous successor node state S at time t does not transmit public opinion information, which is the percentage of S-state nodes with an in-degree of 1 at time t;

[0062] Further classify θ(t) according to its predecessor node status: θ S (t),θ H (t),θ I (t),θ R (t);

[0063] The relationship between these related variables is expressed by the following formula:

[0064]

[0065] In the formula, θ S (t) is the proportion of directed edges that satisfy the definition of θ(t) at time t and whose predecessor node state is S; θ H (t) is the proportion of directed edges that satisfy the definition of θ(t) at time t and whose predecessor node state is H; θ I (t) is the proportion of directed edges that satisfy the definition of θ(t) at time t and whose predecessor node state is I; θ R (t) is the proportion of directed edges that satisfy the definition of θ(t) at time t and whose predecessor node state is R;

[0066] The change of θ over time can be expressed as:

[0067]

[0068] The conditional probability that the predecessor node state of any directed edge is I and the successor node state is S can be expressed as:

[0069] η I (t) = θ I (t) / θ(t):

[0070] The expected number of directed edges from a node in state S with degree (j, k) to its predecessor node in state I is jη I (t):

[0071] Let u jk (t) represents the proportion of nodes with degree (j, k) that still maintain the S state at time t, which satisfies the following differential equation:

[0072]

[0073] To earn points:

[0074]

[0075] In the formula, C1 is the constant obtained by integration.

[0076] And then get:

[0077]

[0078] The constant receive u jk (t) Initial conditions: When false public opinion information begins to spread:

[0079] u jk (0)→1, we know that C1→1:

[0080] u jk The analytical solution of (t) is as follows:

[0081]

[0082] From the above definition of θ(t), we can know that:

[0083]

[0084] The expression for the proportion of S-state nodes with degree (j, k) is:

[0085] S jk (t) = p jk ·u jk (t) = p jk [θ(t)] j .

[0086] Further, a preferred embodiment is provided, wherein the step of calculating the influence of neighboring nodes on the current node and the group of dynamic equations for the propagation of false public opinion information in online social networks comprises:

[0087] For an arbitrary target node, let It represents the probability that the degree of any in-degree neighbor node is (j, k). This probability is also expressed as the probability that the degree of the predecessor node of any directed edge with the target node as the successor node is (j, k);

[0088] The total number of nodes with degree (j, k) in a directed graph network is p jk n, the total number of directed edges from these nodes is p jk ·n·k; the probability that any node has a predecessor node with degree (j, k) It is expressed as:

[0089]

[0090] θ S (t) is derived from the following equation:

[0091]

[0092] The above equation can be further deduced as follows:

[0093] θ S (t) = [∑ j，k k jk (t)] / [∑ j，k <k>n];

[0094] All S jk The ratio of the total out-degree of the type node to the total out-degree of all nodes in the network;

[0095] θ H (t) and θ I (t) is as follows:

[0096]

[0097]

[0098] According to the above propagation rules, θ I The dynamic change of is determined by two parts: the positive term determines the proportion of H-state nodes that spread false public opinion topic information, and the negative term determines (a) a part of I-state nodes become immune state R and (b) successfully infect an S-state node; Satisfy the following formula:

[0099]

[0100] The dynamic equations for the spread of false public opinion information on online social networks are as follows:

[0101]

[0102]

[0103]

[0104]

[0105]

[0106]

[0107]

[0108]

[0109]

[0110] The present invention also proposes a modeling device for a false public opinion information propagation dynamics model, and its technical solution is as follows:

[0111] The device is used to implement the above-mentioned modeling method;

[0112] The device comprises the following modules:

[0113] A module for constructing a directed graph network of an online social network based on the attention mechanism of the online social network;

[0114] The directed graph network includes multiple nodes and multiple edges connecting any two account nodes; the nodes represent independent users in the directed online social network; the edges represent the attention relationship between subscribers and subscribers in the directed online social network; the directed graph network is in, represents the set of nodes; ε represents the set of edges, and

[0115] The directed graph network satisfies the following assumptions:

[0116] Hypothesis 1: Based on the Barabási-Albert model, the topological structure of the directed graph network remains approximately stable during the propagation of false public opinion information in social networks;

[0117] Assumption 2: There are no self-loops or repeated connections between any account nodes in the directed graph network;

[0118] Modules for constructing a propagation dynamics model: The steps for constructing a propagation dynamics model include the following modules:

[0119] A module for defining propagation rules for the spread of false public opinion information;

[0120] Module for obtaining the relationship between the probability of unknown node users becoming propagators and edge characteristics in online social networks:

[0121] A module that calculates the influence of neighboring nodes on the current node and the dynamic equations for the propagation of false public opinion information in online social networks.

[0122] The present invention also proposes a computer device, and its technical solution is as follows:

[0123] A computer device includes: a processor and a memory, wherein the memory is used to store executable instructions of the processor, and the processor is configured to execute the above-mentioned modeling method of the false public opinion information propagation dynamics model by executing the executable instructions.

[0124] The present invention also proposes a computer storage medium, and its technical solution is as follows:

[0125] A computer storage medium stores a computer program, and when the computer program is run, the modeling method of the above-mentioned false public opinion information propagation dynamics model is executed.

[0126] The present invention also proposes a computer program product, and its technical solution is as follows:

[0127] A computer program product includes a computer program / instruction, which, when executed by a processor, implements the steps of the modeling method of the above-mentioned false public opinion information propagation dynamics model.

[0128] The present invention has the following beneficial effects:

[0129] 1. The modeling method of the false public opinion information propagation dynamics model described in the present invention considers the heterogeneity of online social network users and combines the directionality of information propagation to study the related issues such as the false public opinion information propagation process and control strategy in online social networks, and restores the complex online social network structure of false public opinion information propagation. On this basis, the control strategy and practical application of false public opinion information can be further explored, and decision-making recommendations and management opinions can be provided for public opinion controllers to achieve the purpose of ensuring the security of social networks, improving the social network environment and enhancing the efficiency of public opinion control.

[0130] 2. The modeling method of the dissemination dynamics model of false public opinion information described in the present invention accurately summarizes and restores the dissemination law of false public opinion information in online social networks. Since the dissemination of false public opinion information in online social networks has the characteristic of being non-replicable, the research on its dissemination law faces great challenges; the modeling method is based on a large number of relevant literature in the field of dissemination, considering the directionality of information dissemination in online social networks, and attempts to deeply restore the dissemination process of online social network information based on directed complex network theory, and proposes a dissemination dynamics model of false public opinion information in online social networks based on multiple dissemination subjects, multi-layer dissemination carriers and multi-dimensional information based on the SHIR model; the dissemination characteristics of false public opinion information are analyzed from multiple perspectives such as information, users, and platforms, and strive to reproduce its dissemination process and dissemination law.

[0131] 3. The modeling method of the false public opinion information propagation dynamics model described in the present invention can further provide a research basis for early information analysis, the impact of user behavior and platform competition on the propagation process.

[0132] The modeling method and device of the false public opinion information propagation dynamics model described in the present invention are suitable for constructing a false public opinion information propagation dynamics model in a heterogeneous directed online social network. BRIEF DESCRIPTION OF THE DRAWINGS

[0133] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0134] Figure 1 A flow chart of a method for modeling a false public opinion information propagation dynamics model in one embodiment of the present invention;

[0135] Figure 2 In one embodiment of the present invention, an information propagation directed graph is shown;

[0136] Figure 3 In one embodiment of the present invention, a state transition diagram of false public opinion information propagation based on node division;

[0137] Figure 4 In one embodiment of the present invention, a false public opinion information propagation edge state transition diagram based on edge partitioning;

[0138] Figure 5 In one embodiment of the present invention, a Twitter online social network topology diagram;

[0139] Figure 6 In one embodiment of the present invention, the in-degree and out-degree distribution graph of the Twitter online social network platform;

[0140] Figure 7 A Twitter subnetwork degree distribution graph in one embodiment of the present invention;

[0141] Figure 8 A simulation diagram of a model for propagating false public opinion information on an online social network in one embodiment of the present invention;

[0142] Fig. 9 In one embodiment of the present invention, an edge state change graph of an online social network false public opinion information propagation system based on edge partitioning is provided;

[0143] Fig.10 A graph showing the change of Twitter social network node status with in-degree parameters in one embodiment of the present invention;

[0144] Fig.11 A graph showing the change of the Twitter social network node status with the out-degree parameter in one embodiment of the present invention;

[0145] Fig.12 A graph showing the change in the in-degree distribution of nodes in the S state in one embodiment of the present invention;

[0146] Fig.13 A graph showing the change in the in-degree distribution of nodes in the H state in one embodiment of the present invention;

[0147] Fig.14 A graph showing the change in the in-degree distribution of nodes in state I in one embodiment of the present invention;

[0148] Fig.15 This is a graph showing changes in the in-degree distribution of R-state nodes in one embodiment of the present invention. DETAILED DESCRIPTION

[0149] In order to make the technical solutions and advantages of the present invention more clearly described, the specific implementation methods of the present invention will be further described in detail and completely in conjunction with the accompanying drawings. The various implementation methods described below are only part of the preferred solutions of the present invention, rather than all implementation methods; the various implementation methods described below are intended to explain the present invention and cannot be understood as limitations on the present invention; the reasonable combination of technical features defined in the various implementation methods of the present invention, and all other implementation methods obtained by ordinary technicians in the field without creative work based on the implementation methods of the present invention, are within the scope of protection of the present invention.

[0150] In one embodiment, a modeling method for a false public opinion information propagation dynamics model is provided:

[0151] The method comprises the following steps:

[0152] A step for constructing a directed graph network of an online social network according to a follow mechanism of the online social network;

[0153] The (directed) online social network is a directed network consisting of users and unidirectional associations between users; according to its attention mechanism, the online social network is constructed based on a directed graph in graph theory;

[0154] The directed graph network (model) includes a plurality of (account) nodes and a plurality of (follow) edges connecting any two account nodes; the nodes represent independent users in the directed online social network; the edges represent follow relationships between subscribers and subscribers in the directed online social network;

[0155] Constructing directed graph networks using user following relationships in online social networks in, represents the set of nodes; ε represents the set of edges, and

[0156] like Figure 2 As shown in the figure, taking users A, B, C, and D as an example, users B and C have a mutual following relationship. User D follows user C and becomes a fan of C, but user C does not follow user D at the same time. Therefore, all published information between users B and C can be received by each other at the same time, but due to the one-way following relationship between users C and D, user C's information can still be received by D, but C is not aware of all the dynamics of user D.

[0157] The step of constructing a directed graph network of an online social network according to the attention mechanism of the online social network comprises the following steps:

[0158] Steps for node classification:

[0159] The directed graph network satisfies the following assumptions:

[0160] Hypothesis 1: Based on the Barabási-Albert model, the topological structure of the directed graph network remains approximately stable during the propagation of false public opinion information in social networks;

[0161] Assumption 2: There are no self-loops or repeated connections between any account nodes in the directed graph network;

[0162] The account nodes in the directed graph network are divided into four types of nodes (or four states): unknown, hesitant, propagator, and immune;

[0163] The unknown (Susceptible, S): online social network users who have not been exposed to false public opinion information so far, and are still in an unknown state about the public opinion topic, represented by S;

[0164] The hesitating (H): online social network users who have known the false public opinion information but have not yet decided whether to express their opinions or views, represented by H;

[0165] The disseminator (Infected, I): an online social network user who spreads false public opinion information, opinions or views, represented by I;

[0166] The immune persons (Recovered, R) are online social network users who know the false public opinion information but eventually lose interest in it and stop spreading it, represented by R.

[0167] Steps to simplify a directed graph network:

[0168] To simplify the representation of the directed graph network (model), let S represent the proportion of unknown people in the entire group, let H represent the proportion of hesitant people in the entire group, let I represent the proportion of spreaders in the entire group, and let R represent the proportion of immune people in the entire group, and satisfy the following formula:

[0169]

[0170] At time t=0, the group size of communicators in the directed social network used to initiate discussions and spread false public opinion information is assumed to be a neglected value (i.e., can be ignored):

[0171]

[0172] The initial value of the node state ratio of the directed online social network public opinion information dissemination system is approximately:

[0173] S(0)=1, H(0)=I(0)=R(0)=0;

[0174] Steps for degree distribution:

[0175] The degree of a node in a network refers to the number of edges connected to this node;

[0176] In a directed graph network, the in-degree of a node refers to the number of edges entering the node, while the out-degree of a node refers to the number of edges departing from the node.

[0177] In a directed graph network, the in-degree represents the number of users followed by other users, while the out-degree represents the number of other users followed by a user.

[0178] In this directed graph network (or online social network), for any category A∈{S,H,I,M,R}, let A jk (t) represents the proportion of online social network users in state A with in-degree j∈{1, ..., J} and out-degree k∈{1, ..., K} at time t;

[0179] In addition, let p jk The probability that an arbitrary node has degree (j, k) is expressed as follows:

[0180]

[0181] p jk =S jk (t)+H jk (t)+I jk (t)+R jk (t);

[0182] The directed graph network is a random network, and the degree distribution in the random network follows a power-law distribution. The probability of randomly selecting a node with a specific in-degree and out-degree is described by the joint probability distribution F(x, y), and the distribution is expressed as a probability generating function (PGF), and the formula is as follows:

[0183]

[0184] Where x and y are placeholders for constructing the probability generating function;

[0185] In complex networks, average degree is one of the most important properties.

[0186] In a directed graph network, a directed edge is counted as an outgoing edge at its predecessor node and as an incoming edge at its successor node. The total number of outgoing edges and incoming edges is the same, so the average in-degree and out-degree are equal. In this case, the average in-degree of the directed graph network is <j>and average out-degree <k>It is expressed by the following formulas and symbols:

[0187]

[0188]

[0189]

[0190] Where m is the total number of in-degrees in the online social network, which is the same as the total number of out-degrees; v is the total number of nodes in the online social network.

[0191] In this implementation, references to the Barabási-Albert model are:

[0192] Albert R, Barabási AL.Statistical mechanics of complex networks[J].Reviews ofModern Physics, 2002, 74(1):47.

[0193] Steps used for propagation dynamics model building:

[0194] Steps for defining propagation rules for the spread of false public opinion information;

[0195] The propagation rules are as follows:

[0196] Rule 1: Let an S jk Type Node Online social network users receive false public opinion information from one of their neighbors in a dissemination state (i.e., disseminator, I state), which is a Poisson process with a constant rate r;

[0197] When the S jk When a user of type node is exposed to a false public opinion topic, the user's state will immediately change to a hesitant state (i.e., hesitant, H state);

[0198] Rule 2: H jk Type node users need to decide whether to spread false public opinion information;

[0199] Among them, the probability of choosing to spread information related to false public opinion topics is γ, and the user's state will be transformed into a spreading state (i.e., a disseminator, state I);

[0200] The probability of remaining silent and ignoring the public opinion information is 1-γ, and the user's state will be transformed into an immune state (i.e., immune, R state):

[0201] Rule 3: One I jk Type node users may completely lose interest in the public opinion information and transform into an immune state (i.e., immune, R state) at a rate of μ;

[0202] The above node state transition process can be described by the state transition diagram, such as Figure 3 shown.

[0203] Steps for obtaining the relationship between the probability of a node user unknown becoming a propagator and edge characteristics in an online social network:

[0204] In addition, in order to more reasonably represent the heterogeneity of online social network users, it is necessary to further explore the relationship between edges in directed networks:

[0205] Let θ(t) represent the probability that the directed edge with the previous successor node state S at time t does not transmit public opinion information; since the transmission of false public opinion topic information on the directed edge occurs independently following the Poisson process, θ(t) is further interpreted as the percentage of S-state nodes with an in-degree of 1 at time t;

[0206] According to the above definition, θ(t) is further classified according to its predecessor node status: θ S (t),θ H (t),θ I (t),θ R (t);

[0207] The relationship between these related variables is expressed by the following formula:

[0208]

[0209] In the formula, θ S (t) is the proportion of directed edges that satisfy the definition of θ(t) at time t and whose predecessor node state is S; θ H (t) is the proportion of directed edges that satisfy the definition of θ(t) at time t and whose predecessor node state is H; θ I (t) is the proportion of directed edges that satisfy the definition of θ(t) at time t and whose predecessor node state is I; θ R (t) is the proportion of directed edges that satisfy the definition of θ(t) at time t and whose predecessor node state is R;

[0210] Figure 4 Represents the change process of directed edges in online social networks:

[0211] When the state of a predecessor node of a directed edge is I and the node transmits false public opinion information to the successor node in S state through this edge, the directed edge no longer satisfies the definition of θ(t):

[0212] Therefore, the change of θ over time can be expressed as:

[0213]

[0214] According to existing literature, the conditional probability that the predecessor node state is I and the successor node state is S for any directed edge can be expressed as:

[0215] η I (t) = θ I (t) / θ(t):

[0216] Consider a node in state S with degree (j, k), the expected number of directed edges connected to its predecessor node in state I is jη I (t):

[0217] Let u jk (t) represents the proportion of nodes with degree (j, k) that still maintain the S state at time t (i.e., no false public opinion information transmission has occurred), which should satisfy the following differential equation:

[0218]

[0219] Integrating the above equation gives:

[0220]

[0221] In the formula, C1 is the constant obtained by integration.

[0222] And then get:

[0223]

[0224] The constant receive u jk (t) Initial conditions: When false public opinion information begins to spread:

[0225] u jk (0)→1, we know that C1→1:

[0226] Therefore, u jk The analytical solution of (t) is as follows:

[0227]

[0228] From the above definition of θ(t), we can know that:

[0229]

[0230] Therefore, from the definition of S-state nodes and the above equation, the expression for the proportion of S-state nodes with (j, k) degree is:

[0231] S jk (t) = p jk ·u jk (t) = p jk |θ(t)] j .

[0232] So far, the relationship between the probability of a node user changing from S state (i.e., unknown) to I state (disseminator) and edge characteristics in an online social network is obtained.

[0233] Calculate the influence of neighbor nodes on the current node and the steps of the dynamic equations for the propagation of false public opinion information in online social networks:

[0234] In order to derive the dynamic relationship of information propagation in the network, it is necessary to further explore the influence of neighbor nodes on the current node, such as θ I (t) and θ M (t) dynamic changes;

[0235] For an arbitrary target node, let It represents the probability that the degree of any in-degree neighbor node is (j, k). This probability is also expressed as the probability that the degree of the predecessor node of any directed edge with the target node as the successor node is (j,).

[0236] In addition, the total number of nodes with degree (j, k) in the directed graph network is p jk n, so the total number of directed edges from these nodes is p jk ·n·k; then, the probability that any node has a predecessor node with degree (j, k) is It is expressed as:

[0237]

[0238] According to the above definition, θ S (t) is derived from the following equation:

[0239]

[0240] The above equation can be further deduced as follows:

[0241] θ S (t) = [∑ j，k k jk (t)] / [∑ j，k <k>n];

[0242] That is, all S jk The ratio of the total out-degree of the type node to the total out-degree of all nodes in the network;

[0243] θ H (t) and θ I (t) A similar method is used to derive as follows:

[0244]

[0245]

[0246] According to the above propagation rules, the dynamic change of,θ,is determined by two parts: the positive term, the latest decision of the proportion of H-state nodes that spread false public opinion topic related information, and the negative term, (a) a part of I-state nodes become immune state R and (b) successfully infect an S-state node; Satisfy the following formula:

[0247]

[0248] According to θ(t), θ S (t),θ H (t),θ I (t),θ R With the definition of (t) and the above equation, the dynamic equations for the spread of false public opinion information on online social networks are as follows:

[0249]

[0250]

[0251]

[0252]

[0253]

[0254]

[0255]

[0256]

[0257]

[0258] It should be noted that with the continuous rise of mobile smart terminals, social networks have invisibly penetrated into daily life, and traditional social networks such as mobile phone address books and emails have rapidly evolved into various online social networks. In this process, the network structure and nature of social networks have also undergone tremendous changes.

[0259] It should be noted that in recent years, with the increasing popularity of the Internet and mobile smart terminals, the way people exchange information has gradually changed from offline face-to-face interaction to online virtual social interaction, thus forming an online social network with the Internet as the medium. Among them, each node represents a registered user in the network, and the edge represents the virtual connection between nodes. The emergence of online social networks has made the connection between people in the social network closer, and the speed of information dissemination has been greatly improved.

[0260] Based on the context of online social networks, this implementation method defines online social network false public opinion information as "false, harmful, erroneous or even reactionary information, opinions or views related to a public event that is widely concerned by the general public and that are published and disseminated through online social networks, and these comments often have certain tendencies and influences."

[0261] To further refine the above definition, the public events that are widely concerned by the public and focused on by this implementation method refer to public safety events that occur suddenly and cause or may cause major casualties, property losses and major social impacts, such as public health events, emergency disaster safety events, etc. In addition, the "false" in the definition of false public opinion information on online social networks is mainly judged through the clarification information content released by public opinion controllers, various types of false information released in real time by the network platform management center, etc., to identify the authenticity of related topics of public events.

[0262] Based on the above definition, the main characteristics of false public opinion information on online social networks can be summarized as follows: 1) Deliberately highlighting controversial elements. Controversial elements are most likely to attract the attention of users in online social networks. Therefore, some users and even platforms tend to highlight certain controversial elements when pushing messages to gain popularity; 2) Taking public information out of context. Online social network users have different prior knowledge of a public opinion event, which causes users with relatively less prior knowledge to have an inadequate understanding of the public opinion controller’s public information, and then distort it and publish opinions or views that are taken out of context; 3) Fabricating unconfirmed rumors. Some users will use sensational headlines to guide public sentiment based on the public opinion points that the general public is concerned about.

[0263] It should be noted that due to the non-replicability of the spread of false public opinion information on online social networks, most current studies follow the research steps of "factor analysis - model construction - theoretical derivation - simulation experiment". Through the review of existing literature, scholars' current research on the spread of false public opinion information mainly focuses on the analysis of macro-evolutionary laws and micro-theoretical modeling. Among them, theoretical modeling research can be summarized into three categories, namely complex network propagation models, infectious disease propagation models, and game theory models. They are introduced below.

[0264] For complex network communication models: The assumption of uniformly mixed population in the basic epidemic dynamics model is a limitation of public opinion information dissemination research, and it cannot completely restore the dissemination process of false public opinion information in online social networks. In order to establish a dissemination model that is more in line with the characteristics of online social networks, technicians in this field introduced complex networks with different topological structures into the network, and gradually transferred from static network topological structures to dynamic network topological structures. Based on empirical analysis, the corresponding change laws of its topological structure in the dynamic evolution of online social networks were discussed. However, there are many factors that affect the evolution of online social networks, and most research methods in this field are based on empirical research, which means that the universality of research on the laws of changes in topological properties in the dynamic evolution of networks still needs to be further improved. Limited by this, there are also certain difficulties and limitations in the research and development of dynamic network and multi-layer network information dissemination based on the laws of changes in topological structures.

[0265] For infectious disease propagation model: it is the most widely used propagation model in modeling the propagation of false public opinion information in online social networks. Based on the DK model, technicians in this field have further expanded and improved it from multiple aspects such as user state division, heterogeneous nodes, and user state transition probability. Most models that use user state division are based on node division, and do not consider the impact of the degree distribution of the network on the information propagation process. Obviously, most online social networks do not meet such strict assumptions. Most models that use user heterogeneity expression use several undirected networks based on ordinary differential equations based on edge partitioning theory, and there is almost no literature based on directed network exploration. In addition to the probability generating function method, energy calculation formulas, clustering coefficients, semi-directed networks and other methods have also been proposed to try to solve the information propagation problem of directed networks, but the effect is still not good. Most models that use user state transition probability take the internal structure of the network and the external environment of the network as influencing factors, and basically cover the user state transition probability expression method considering the characteristics of the online social network platform itself. However, the spread of public opinion information is not only affected by the platform's own structure and environment, but also closely related to internal factors such as user behavior.

[0266] For game theory models: The most commonly used method to study the impact of user behavior on state transition probability and even the law of information dissemination is game theory. Game theory models reveal the laws and mechanisms of public opinion information dissemination in online social networks by studying the strategic choices, information dissemination paths, and interactions between participating entities in the game. However, most of the literature focusing on online social network users only treats users as a single concept in the game, and there is a relative lack of research that introduces the user's own psychological state and the superimposed influence mechanism of network user groups on information dissemination into the dissemination of public opinion information.

[0267] It should be noted that most behaviors in online social networks can be summarized as communication behaviors that follow certain rules. How to describe the overall process and specific characteristics of the communication behavior, and then explore effective methods and specific measures to control the behavior, is currently a hot topic for most scholars. In recent years, most scholars have restored, simulated and demonstrated the communication behavior in online social networks by building models, and analyzed some potential consensus and rules of the communication behavior by setting and adjusting certain parameters of the communication behavior, thereby providing some reference opinions for the control of false public opinion information. Based on the similarities between the spread of information between people and the spread of infectious diseases, most of the current modeling research on information has borrowed from the infectious disease transmission dynamics model.

[0268] Most infectious disease transmission dynamics models ignore the specific changes in population size caused by objective factors such as birth, death, and migration in reality. Instead, the model only considers the type of state the population belongs to and the transition changes between states.

[0269] At present, according to different transmission states and transmission processes, the classic infectious disease transmission dynamics model can be divided into susceptible state-infected state (SI), susceptible state-infected state-suspected state (SIS), susceptible state-infected state-recovered state (SIR), susceptible state-latent state-infected state-recovered state (SEIR), susceptible state-hesitating state-infected state-recovered state (SHIR) transmission models.

[0270] Most of the above-mentioned classic infectious disease transmission dynamics models still have certain limitations in their application in the field of research on the spread of false public opinion information.

[0271] The repair time scale in the propagation process described by the SI propagation model will be much larger than the propagation time scale, which means that those people in online social networks who are exposed to false public opinion information but do not choose to spread it are not accurately restored.

[0272] Although the SIS propagation model has been further improved on the basis of the SI model, the dynamic propagation process constructed by the SIS propagation model is reversible, that is, as long as the current propagation rate is always greater than the propagation threshold of public opinion information in the network, there will always be users in the infected state I in the online social network, which is inconsistent with the spread of false public opinion information in reality; furthermore, after being infected by false public opinion information, users become informed of the incident. Even if they have recovered to a state without the ability to propagate later, they should still be distinguished from the ignorant people represented by the initial susceptible state S.

[0273] The SIR propagation model further compensates for this limitation on the basis of the SIS model, that is, the propagation process of false public opinion information in this propagation model will eventually evolve into a situation where there are no longer any infected users. It is not difficult to find that the SIR propagation model is very similar to the propagation of false public opinion information in terms of population classification and propagation rules. Therefore, most scholars use the SIR model as a theoretical basis, and further expand and improve it to adapt to propagation problems in other fields. However, the SIR propagation model still only restores the propagation of public opinion information on the network from a macro perspective. Unlike the passive acceptance and compulsory propagation of infectious diseases, whether network users accept and spread information is selective. Therefore, the degree of restoration of the propagation of false public opinion information in online social networks at the micro user level still needs to be further improved.

[0274] The SHIR propagation model makes up for the limitations of the SIR model in user selection, making great progress in the study of the spread of false public opinion information. However, the SHIR propagation model is still limited by the basic assumption of an undirected and homogeneous network in the classic infectious disease model, which makes it difficult to accurately describe the directionality of connections between network users, the unidirectionality of information dissemination, and the heterogeneity of network users themselves.

[0275] In summary, the characteristics of the classic infectious disease dynamics model are very similar to the pattern of false public opinion propagation in online social networks, but the online public opinion propagation model that takes into account the unidirectional nature of information propagation and user heterogeneity still needs to be explored.

[0276] It should be noted that in the real world, the classical infectious disease dynamics model can no longer meet the research on the spread of false public opinion information in online social networks. Due to the directionality of the connection between accounts in online social networks (subscription and follow-up relationships between accounts) and heterogeneity (the number of fans of different accounts is different), the use of homogeneous undirected networks for description has certain limitations. Therefore, it is crucial to improve the accuracy and reliability of the model's description of the spread of false public opinion information in online social networks. The model described in this embodiment explores the basic state division rules of users in the process of spreading false public opinion information in online social networks, and uses edge partitioning theory to construct a basic information propagation dynamics model, which can improve the accuracy and reliability of the description of the spread of false public opinion information in online social networks.

[0277] It should be noted that in real-world applications, there are certain deficiencies in using the classic infectious disease dynamics model to describe the typical spread of false public opinion information on online social networks. On the one hand, the connection between online social network accounts is directional, that is, the subscription and attention relationship between accounts, and this phenomenon cannot be simply represented by an undirected graph; on the other hand, the connection between online social network accounts is heterogeneous, and the number of fans of different accounts is different. Using a homogeneous network to describe it has certain limitations. Therefore, in order to improve the accuracy and reliability of the model's description of online social networks, this implementation method considers the directional nature of information dissemination in online social networks, that is, only the fan group that pays attention to the account that spreads information can receive information in a unidirectional manner, while the user group that the account that spreads information itself pays attention to cannot receive information. At the same time, the number of fans and attention of online social network accounts is further refined to express its heterogeneity. This implementation method takes online social networks as the research object, and aims to demonstrate the feasibility and accuracy of the restoration of false public opinion information dissemination by directed online social networks based on user heterogeneity. Focusing on the two core issues of information dissemination process and empirical comparative analysis in social networks, this implementation method constructs a propagation dynamics model of false public opinion information in online social networks, and uses network structure data to restore an online social network (such as Twitter) to verify the effectiveness of the model; on this basis, the dynamics model is compared with the traditional infectious disease model to identify the key network structure factors that affect the spread of false public opinion information.

[0278] It should be noted that there is currently a lack of false public opinion information dissemination models that consider the directionality and heterogeneity of connections between accounts. The academic community has fully recognized the differences between online social network information and infectious disease transmission, and has expanded more population states that are suitable for the information dissemination process based on the classic infectious disease transmission model. However, most of the current research on the information dissemination process and control strategies of online social networks is based on the rough assumption that users who take the same behavior at a certain moment are identical, resulting in a certain degree of deviation in the restored information dissemination process; compared with undirected networks, the asymmetric information dissemination of directed networks is more realistic and universal, but it is rarely considered in existing studies. How to break through the assumptions of a single online social network user group and undirected information dissemination, and explore the dissemination, judgment and control strategies of false public opinion information in online social networks is an important issue that needs to be solved. In addition, most studies have ignored the process of online social network users seeking and discriminating public opinion information. The few literatures that focus on network users only use it as a single concept in the game, and do not introduce the user's own psychological state and the superimposed influence mechanism of the user group on information dissemination into the false public opinion information dissemination model. The exploration of the dynamic interaction phenomenon between individuals and groups still needs to be strengthened.

[0279] It should be noted that this implementation method is mainly based on basic theoretical knowledge such as complex network theory, game theory and optimization theory, and applies complex network analysis, population dynamics and system stability analysis methods to study the law of dissemination of false public opinion information on online social networks. The details are as follows:

[0280] (1) Complex network analysis methods

[0281] Compared with the material transmission methods such as traditional media information and infectious diseases, the transmission of false public opinion information in online social networks is essentially a new "heterogeneous + directed" transmission mode. In this transmission mode, users are no longer single individuals, but show heterogeneity by following other users; at the same time, this attention mode also gives information transmission directionality. The complex network of information transmission is evolved from network users and their associations. This implementation method uses complex network analysis methods to restore the topological structure of online social networks, and further explores the impact of network properties such as degree distribution on transmission.

[0282] (2) Population dynamics method

[0283] Population dynamics is a quantitative research method for calculating the dynamic changes in the number of individuals. The main idea is to construct a differential dynamics system to simulate the change process of the number of individuals. The infectious disease model is a classic model constructed using the population dynamics method. This implementation method uses this quantitative research method to model the process of false public opinion information dissemination from the perspective of online social network users and account status changes, and analyzes the specific impact of users and platforms on the dissemination of this information.

[0284] (3) Game analysis method

[0285] Equilibrium analysis is an analytical method that seeks to achieve equilibrium conditions under the assumption that various economic variables and their correlations are determined. This embodiment uses the equilibrium analysis method to analyze and discuss the equilibrium solution of the communication game of hesitant users in online social networks and the zero-sum differential game between platforms.

[0286] In this embodiment, the modeling method is further improved based on the SHIR classic propagation model, taking into account the directionality of information propagation in online social networks and the heterogeneity of users, and constructing a directed complex network and heterogeneous public opinion information propagation dynamics model.

[0287] In this embodiment, the modeling method studies the directed propagation process of false public opinion information on the Internet from the perspective of the diversity of online social network users, and constructs a corresponding dynamic model.

[0288] In this embodiment, the modeling method applies edge partitioning theory to construct a directed complex network to address the problem of unidirectional network information propagation, and further explores the propagation dynamics model of false public opinion information.

[0289] In this embodiment, the modeling method considers the heterogeneity of online social network users and combines the directionality of information dissemination to study related issues such as the dissemination process and control strategy of false public opinion information in online social networks, restores the complex online social network structure of false public opinion information dissemination, and on this basis can further explore the control strategy and practical application of false public opinion information, provide decision-making recommendations and management opinions for public opinion controllers, so as to achieve the purpose of ensuring the security of social networks, improving the social network environment and enhancing the efficiency of public opinion control.

[0290] In this embodiment, the modeling method can further provide a research basis for early information analysis, user behavior and the impact of platform competition on the dissemination process.

[0291] In this embodiment, the modeling method is based on the theory of directed complex networks to construct a directed network architecture of online social networks to restore the directional characteristics of public opinion information dissemination; on this basis, the edge partitioning theory is used to improve the heterogeneous expression of users, and a (online social network) false public opinion information dissemination dynamics model based on the SHIR model (Susceptible-Hesitating-Infected-Recovered Model) is proposed to accurately restore the heterogeneous structure of online social network users and the directed information dissemination process.

[0292] In this embodiment, the modeling method introduces the directionality and heterogeneity characteristics of the connections between online social network accounts, migrates the classic infectious disease dynamics model to the field of online social network research, and uses edge partitioning theory to correspond and refine the node's own state, connection method and information propagation law to construct a (heterogeneous directed online social network) false public opinion information propagation dynamics model.

[0293] In this embodiment, the modeling method accurately summarizes and restores the propagation law of false public opinion information in online social networks. Since the propagation of false public opinion information in online social networks is non-replicable, the research on its propagation law faces great challenges. The modeling method is based on a large number of relevant literature in the field of communication, considering the directionality of information propagation in online social networks, and attempts to deeply restore the propagation process of online social network information based on directed complex network theory, and proposes a dynamic model of false public opinion information propagation in online social networks based on multiple propagation subjects, multi-layer propagation carriers and multi-dimensional information based on the SHIR model; analyzes the propagation characteristics of false public opinion information from multiple perspectives such as information, users, and platforms, and strives to reproduce its propagation process and propagation law.

[0294] In this embodiment, the modeling method takes into account the directional and heterogeneous characteristics of the connection between online social network accounts, studies the directional propagation law of false public opinion information in online social networks, and demonstrates the feasibility and accuracy of the restoration of false public opinion information propagation in real networks by directional online social networks based on user heterogeneity. The classical infectious disease dynamics model is transferred to the research field of online social networks, the basic states of users in the process of false public opinion information propagation in online social networks are divided, and the rules for the propagation of false public opinion information and the fixed state transition probabilities between user states are defined. The edge partitioning theory is used to construct a dynamics model for the propagation of false public opinion information in online social networks.

[0295] In this embodiment, for false public opinion information on online social networks, how should public opinion controllers accurately restore its propagation process, that is, how to determine the structure of the propagation network? What are the key factors that affect the propagation of false public opinion information? What is the specific propagation process? The above problems are solved by the modeling method.

[0296] In addition, in one implementation, a specific example is provided:

[0297] Analysis of the dissemination process of false public opinion information in online social networks

[0298] This section uses an example of an online social network (Twitter) to compare the online social network model that introduces directionality and heterogeneity with the online social network model of traditional research, and further verifies the accuracy and effectiveness of the directed online social network structure model based on edge partitioning theory described in the above implementation method.

[0299] Firstly, we restore Twitter online social network data and analyze the Twitter network structure;

[0300] Secondly, the parameter settings of the example are described;

[0301] Secondly, the dissemination of false public opinion information in directed networks based on edge partitioning theory is compared with the dissemination of traditional network information based on node partitioning.

[0302] Finally, the example results are analyzed and summarized.

[0303] Analyzing Twitter network structure:

[0304] In order to comprehensively analyze the network structure characteristics of online social networks, Twitter, a well-known foreign online social network platform, was selected as a sample. The Twitter user relationship dataset was published by the Data Mining and Machine Learning Laboratory of Arizona State University, and specifically includes CSV format files of nodes and edges.

[0305] According to statistics, the Twitter user relationship dataset contains 11,316,811 node users and 85,331,846 user relationship directed edges. Due to the large number of node users and edges in the dataset, it is difficult to generate a network graph structure using all the nodes.

[0306] Therefore, a node in the Twitter user relationship dataset was randomly selected and a connected subgraph was generated based on the node to ensure that the selected sample could restore the overall structure of the Twitter online social network as much as possible. The number of nodes in the subgraph was 10,000.

[0307] Figure 5 It shows the topology of Twitter online social network.

[0308] The details are as follows:

[0309] The network scale structure type is generally determined by the clustering coefficient and the average path length.

[0310] The clustering coefficient indicates the degree of clustering between nodes in the network, that is, the closeness between any node and its neighboring nodes. The clustering coefficient C of the network is:

[0311]

[0312] Where:

[0313] C i ——The clustering coefficient of any node i in the network;

[0314] v i ——The number of nodes directly connected to any node in the network;

[0315] e i ——The number of edges between any node in the network and its directly connected nodes.

[0316] From the equation, we can see that the clustering coefficient of the Twitter social network generated by this Twitter user relationship dataset is 0.017. The average path length represents the average number of nodes in the shortest relationship chain between any two nodes in the network. Further, we can see that the average path length of the Twitter directed network is 4.17, which is lower than Facebook (4.74) and MSN (6.6).

[0317] The analysis found that the network has a high clustering coefficient and a small average path length, so the Twitter online social network can be classified as a small-world network. Although only a small number of shortcuts are formed in the network, the time required for public opinion information to propagate in the small-world network model is very close to the time it takes for such information to propagate in a random graph, that is, infectious diseases spread rapidly in a small-world network. By analogizing this small-world characteristic to online social networks, various types of false public opinion information spread rapidly in social networks such as Twitter, and how public opinion controllers can detect and control it has become a current concern.

[0318] The above content analyzes the topological properties of the Twitter social network and gives a general description of the network structure from a macro level.

[0319] On this basis, we further focus on the micro-level such as the internal structure of the network and the degree distribution of network nodes, and analyze the Twitter social network from five network topological properties: network degree, degree distribution, network density, reciprocity and assortativity.

[0320] Network node degree:

[0321] From the definition of degree, we know that the degree is the number of edges connected to the target node, and the in-degree and out-degree represent the number of edges entering or departing from the node respectively. These indicators can intuitively reflect the important properties of nodes in complex networks. In addition, the point in the network that is not connected to any edge has a degree of 0 and is called an isolated point in the complex network. In network structure data, the degree can be expressed by the adjacency matrix Z = (z ab ) v×v The specific formula for calculation is as follows:

[0322]

[0323] Where:

[0324] z ab , z ba ——the network elements in its adjacency matrix;

[0325] a, b – placeholders for building the adjacency matrix elements.

[0326] The above equations can be used to calculate the topological properties of nodes such as in-degree and out-degree in the Twitter user relationship dataset. The specific results are shown in Table 1.

[0327] Table 1 Statistics of user degree of Twitter online social network nodes

[0328]

[0329] It can be seen that the maximum value of the node degree in the Twitter social network is 10300, which is much different from the minimum value of the node degree, 1, which indicates that the network is a heterogeneous network; the average value of the total degree is 60, which further indicates that the degree of most nodes in the network is small, and only a few nodes have relatively large degrees. Therefore, the Twitter social network has the characteristics of a scale-free network.

[0330] In addition, the maximum value of the node in-degree in the data set is 3370 and the minimum value is 0, indicating that the behavior of users following others in the network is quite different. Some node users do not follow other accounts at all, and their use of the Twitter platform is still limited to searching for information and viewing hot issues. However, a small number of node users follow a large number of accounts. When using the Twitter platform, they tend to receive information from the accounts they follow in the first place, and have a psychological state of continuous attention to a topic or interest. At the same time, online social networks are directed complex networks, and information flows from the accounts they follow to the node users whose relationships are unidirectional. This shows that if a node user follows more accounts, the more channels he can obtain topic information, and the higher the efficiency of obtaining information. Users who follow fewer accounts cannot obtain updates and push notifications of topic information in the first place, so their efficiency in obtaining information is relatively low.

[0331] As shown in Table 1, the maximum value of the node out-degree in the Twitter user relationship data set is 8660, indicating that there are "super-spreading nodes" in the network, that is, authoritative users. When the number of followers of a Twitter user reaches a certain considerable number, it has a great influence on the spread of information in the network. The information released by this user will not only have a leading role for ordinary users, but may also promote the development of network events. The emergence of "super-spreading nodes" is no longer limited to satisfying the self-need to express opinions, but has become an important "opinion vane" for online social networks and even social groups, with far-reaching economic and social value.

[0332] Network degree distribution:

[0333] The above quantitative indicators describe in detail the size of each node degree in the Twitter user relationship dataset, and provide a macroscopic understanding of the possible distribution of node degrees. On this basis, the degrees of all nodes in the Twitter user relationship dataset are sorted out to form a more intuitive Twitter degree distribution, such as Figure 6 As shown:

[0334] Figure 6 The assumption that the degree distribution of online social networks conforms to the power law distribution is verified. In the Twitter social network, the in-degree (number of followers) and out-degree (number of fans) are relatively balanced. Accounts with fewer followers or fans occupy the vast majority of the network, and their number is about 10,000 times that of accounts with more than 10,000 fans.

[0335] In addition, it can be seen from the figure that when the degree is small, the number of in-degrees in the network is generally greater than the number of out-degrees. This is due to the operating mechanism of online social networks. That is, although the popularity (number of fans) of an account is relatively low, it can still follow more accounts to obtain information. This is also the mode in which most network users use online social networking platforms. However, users in online social networking platforms do not follow other accounts endlessly. When the popularity of an account gradually increases, the number of its fans, that is, the out-degree, increases and gradually exceeds the number of followers of the user. This inflection point can be seen from Figure 3 Finally, when the user gradually grows into a "star account", the number of his fans will far exceed the accounts he follows.

[0336] Network Density:

[0337] Network density can be used to characterize the density of the edges between nodes in the network. It is often used to measure the density and evolution trend of social relationships in social network research. It is defined as the ratio of the actual number of edges in the network to the upper limit of the number of edges that can be accommodated. In addition, the value range of network density is [0,1]. When the network density is 1, the network is fully connected, that is, each node in the network is connected to each other; on the contrary, when the network density is 0, each node in the network is isolated and there is no edge relationship. According to the definition of m and v, a network G with v nodes and m edges, its network density d(G) should satisfy the following equation:

[0338]

[0339] The above equation can be used to calculate the network density of the Twitter user relationship dataset to be 0.00000133. The network density of the Twitter social network is low, indicating that the relationship between users in the network is not close enough, that is, users are not used to following accounts to receive information, and there are not many connections between accounts.

[0340] Reciprocity:

[0341] In actual network problems, people are interested in determining the possibility of double links (in opposite directions) between pairs of vertices (nodes). The concept of reciprocity can measure the degree of mutual connection between two nodes in a directed network. If node user a follows node user b, and node user b follows node user a back, then it is considered that the relationship between the two is reciprocal. The introduction of reciprocity can measure the closeness of interaction between users in online social networks, and it can also measure the error caused by ignoring the directed nature of the network. The reciprocity coefficient φ is used to quantify reciprocity, which is defined as the probability that there is a directed edge from node b to node a when any directed edge from node a to node b is selected in a directed network. It can be seen that its mathematical expression is:

[0342]

[0343] Where:

[0344] m d ——The number of reverse edges in the network;

[0345] z ab z ba ——The product of the elements of the adjacency matrix in the network. The product is 1 if and only if there is a bidirectional directed edge from node a to node b, otherwise it is 0.

[0346] According to this equation, we can calculate that the reciprocity coefficient of the network generated by the Twitter user relationship dataset is 0.052. Compared with livejournal and wiki-Talk, the reciprocity of this network is relatively lower, the degree of interaction between users in the network is reduced, and the flow of information between network users is slower than that of smaller-scale online social networks. However, based on its large-scale network structure, the spread and diffusion of false public opinion information cannot be ignored.

[0347] Network Assortativity:

[0348] Assortativity is mainly used to examine whether nodes with similar degree values ​​tend to connect to each other. This quantitative indicator can effectively reflect the strength of association between node users. In general, if nodes with larger degrees in the network tend to connect to nodes with similar degrees, the degree of the network is said to be positively correlated, that is, the network is an assortative network; if nodes with larger degrees in the network tend to connect to nodes with smaller degrees, the degree of the network is said to be negatively correlated, that is, the network is an incompatible network. Network assortativity can be quantified by the degree of the nodes. Specifically, the assortativity coefficient κ can be calculated from the Pearson product-moment correlation coefficient (PCCs), and the formula is as follows:

[0349]

[0350] Where:

[0351] c i , d i ——The degree of the nodes at both ends of the i-th edge in the network.

[0352] In social networks, the edges between nodes represent more of a social relationship, such as family, friends, colleagues, classmates, etc. In a social environment, people are always more inclined to associate with people who are similar to themselves in many aspects, which makes the real social network generally appear as a homogeneous network. In addition, unlike social networks, there are still complex networks in other fields such as e-mail and biology in the real world, most of which show a heterogeneous network phenomenon, as shown in Table 2 below.

[0353] Table 2. The matching coefficients of multiple types of networks

[0354]

[0355] Due to the strong anonymity, strong interactivity, and wide participation of online social networks, the objective hierarchical differences in social networks are eliminated, allowing anyone to "close the distance" to "super users" through behaviors such as following, posting, forwarding, and commenting, thereby establishing a certain connection. Therefore, online social networks generally appear as heterogeneous networks. For example, the homogeneity coefficients of Renren and Cyworld are -0.0036 and -0.13 respectively. It can be obtained that the network generated by the Twitter user relationship data set has a homogeneity coefficient of -0.102, which belongs to a heterogeneous network. This once again verifies that in online social networks such as Twitter, user groups tend to associate with accounts that receive more attention.

[0356] The following is an explanation of the example settings.

[0357] Example data description:

[0358] Depend on Figure 3 It can be seen that accounts with fewer in-degrees and out-degrees occupy the vast majority of the network, while accounts with larger in-degrees and out-degrees are very rare and can be ignored in statistics. Therefore, taking the number of in-degrees and out-degrees in the network as 300 as the dividing line, a node within this range is randomly selected and a connected subgraph is generated based on this node for the analysis of the dissemination process of false public opinion information. The number of nodes in this subgraph is 500,000. Figure 7 This shows the degree distribution of the connected subgraph:

[0359] and Figure 3 In contrast, this subnetwork basically restores the degree distribution of the Twitter user relationship dataset and consists of Figure 7 (b) It can be seen that most of the nodes in the connected subnetwork are still distributed within the range of in-degree and out-degree of 50, and the number of followers in an account is generally greater than the number of fans.

[0360] Parameter settings for the example:

[0361] This example is based on an example of a popular public opinion topic that is widely spread in social networks (such as pneumonia caused by the new coronavirus). The example assumes that the public opinion topic relies on a random connected subgraph with 500,000 nodes in the above network. Through simulation, the online social network public opinion propagation model based on node partitioning and edge partitioning is compared to analyze the impact of degree distribution on the propagation process and node status. The specific parameters of the numerical experiment are shown in Table 3.

[0362] Table 3 Parameter values ​​of the examples

[0363]

[0364] The following is an analysis of the propagation process:

[0365] The false public opinion propagation models in online social networks based on node partitioning and edge partitioning are compared, the accuracy of the proposed directed heterogeneous online social network false public opinion information propagation model is analyzed, and on this basis, the impact of degree distribution on the propagation process and node status is explored.

[0366] Comparison of the spread of false public opinion information in online social networks by node division and edge division:

[0367] The example simulates the false public opinion information propagation system of online social network based on node partitioning and edge partitioning. In the Twitter social network with connected subgraphs, the node states of the two systems change over time as shown in the following figure. Figure 8 (a) and 8(b).

[0368] from Figure 8 (a) It can be seen that in the online social network false public opinion information propagation system based on node division, almost all unknown user groups in the network are infected by false public opinion information, and the infection speed is relatively fast. Among them, the decline rate of S-state users is slow in the initial stage, but with the increasing number of I-state users, the decline rate increases and the curve changes steeply, approaching 0 at 400 time steps; the change trends of H-state and I-state users are very similar, both of which increase rapidly in the initial stage, and then decline rapidly to 0 after reaching the peak, and the increase rate of H-state users in the initial stage is greater than that of I-state users. This is because after being infected by false public opinion information, S-state users need to enter the H-state first and then transform into the I-state at a certain proportion. The peak value of the I-state is higher than that of the H-state, which indicates that users in online social networks pay attention to this type of public opinion information for a long time and are not easy to lose interest; the change trend of R-state users is exactly the opposite of that of S-state users. In the initial stage, due to the scarcity of I-state users, they increase slowly. In the middle stage, as the number of I-state users continues to increase, the group of users in this state expands rapidly until all users in the network are immune. This reflects that false public opinion information has been fully disseminated in the dynamic model based on node division, that is, all users in the online social network are aware of this type of public opinion information. However, this phenomenon does not exist in real online social networks.

[0369] The node state changes of the online social network false public opinion information propagation system based on edge partitioning are as follows: Figure 8 (b) shows the evolution trend of H state and I state. Figure 8 (a) is similar, that is, it rises rapidly to a peak and then decreases rapidly until it disappears; Figure 8 (a) The difference is that the proportion of S-state users decreases rapidly in the initial stage and remains at 0.35 at 400 time steps. In contrast to S-state users, the proportion of R-state users increases rapidly in the initial stage. This is because the users in the spreading state in the network continue to lose interest in false public opinion information and enter an immune state. However, the growth rate of the R-state gradually decreases and finally stabilizes at 0.65 in the later stage because no new spreading users are generated in the online social network.

[0370] Figure 8 The changing trends of the states of each node in the dynamic system represented by (a) and 8(b) are roughly similar, but it is worth noting that in the false public opinion information propagation system based on node division, the R state ratio is 1, indicating that all users in the network are aware of the false public opinion information, while the R state ratio in the public opinion information propagation system based on edge division is not 1, indicating that some users in the system have not received such public opinion information. This phenomenon is consistent with Zanette's research results, that is, the spread of information in the network has certain scale limitations, and it is impossible for all users in the network to be infected.

[0371] This example further analyzes the time-varying edge states of the online social network false public opinion information dissemination system based on edge partitioning. Fig. 9 shown.

[0372] although Fig. 9 and Figure 8 The overall evolution trend of the states in (b) is similar, but the specific change trends are different. Figure 4 ,θ I The increase in propagation time is weak because only θ H Convert to θ I This is the only way, and θ I It can be transformed into θ R Combined with the two transformation modes of directed edge with predecessor node state H and successor node state I, combined with the parameter settings of r, γ and μ in the example, it can be considered that θ I The input of is less than the output, so it has almost no growth trend during the propagation process. On the contrary, although θ H It can be obtained by θ S is transformed, and it can itself be transformed into θ I and θ R , whose input and output paths are related to θ I Similar, but due to the example parameters, the input is greater than the output in the initial stage, so a peak is formed in a short time, but as θ S The decrease of θ H The ratio of θ gradually decreases until it reaches zero. It is worth noting that after the propagation is completed, the θ S With θ R The sum of is not θ, which just verifies the following equation:

[0373]

[0374] The impact of degree distribution on the propagation process:

[0375] Figure 8 and Fig. 9 The changes of node states in Twitter social network are described from a macro perspective. However, the most important user heterogeneity and its sensitivity analysis on propagation, that is, the impact of in-degree and out-degree of online social network on the propagation of false public opinion information on the network, needs further exploration.

[0376] and Fig.10 It reflects the change of Twitter social network node status with the in-degree parameter. Among them, the number of S-state users decreases rapidly during the whole propagation process and the speed is positively correlated with the in-degree. Fig.10 (f) As the spread of public opinion ends, some S-status users still remain in the Twitter social network. Figure 7 (a) shows that Twitter social network users with fewer in-degrees account for a higher percentage of the total users, so there are still more users with fewer in-degrees retained. Fig.10 The state shown in (b) was in a continuous growth trend before. When H-state users continuously transformed into I or R-state users, the proportion of H-state users in the Twitter social network gradually decreased until they basically disappeared at the end of the propagation. In the entire propagation process, the proportion of H-state users with different in-degrees was roughly the same. The growth time node of I-state users was after the peak of H-state users and the proportion of I-state users was positively correlated with their in-degree. When the number of I-state users reached the peak, it began to gradually decrease until all of them were transformed into R-state users. The R-state users showed a slow and continuous increase in the process of false public opinion information propagation in the entire Twitter social network.

[0377] and Fig.10 In contrast, the Twitter social network user status changes with the out-degree parameter as shown in Fig.11 As shown. Overall, the curve of Twitter social network user status changing with in-degree is relatively smoother than the curve of out-degree. This is because the random connected subgraph with 500,000 network nodes that the example relies on has an uneven distribution of sampling nodes during the collection process. The figure shows that the random connected subgraph of the example has better network node in-degree collection than out-degree. In addition, the trend of the out-degree proportion of users in S, H, I, and R states is roughly similar to that of the in-degree, that is, the number of S-state users decreases rapidly, the proportion of H-state users continues to grow to the maximum value and then gradually decreases until it disappears, the peak of the I-state user proportion is later than that of the H-state and the duration is relatively long, and the R-state users increase slowly throughout the process. But by contrast Fig.10 (f) Fig.11 (f) It can be seen that when the spread of false public opinion information on Twitter tends to be stable at the end, the out-degree of users in S state, I state and R state is significantly higher than the in-degree, which means that the in-degree of users in the Twitter social network has a greater impact on the spread of false public opinion information than the out-degree.

[0378] The impact of degree distribution on node status:

[0379] Focusing on the degree distribution of node states, we analyze how the degree distribution of S, H, I, and R states changes over time.

[0380] The distribution of S state changes as Fig.12 As shown in the figure, at the initial moment, the number of followers of Twitter social network users is generally greater than the number of fans. On the whole, the number of S-state users decreases as the spread of false public opinion information deepens, among which the S-state nodes with large in-degree decrease first (converted to H-state). This is because the more accounts online social network users follow, the wider the source of public opinion information they obtain, and the greater the probability of converting from S-state to H-state; while the relatively insignificant change of S-state nodes with large out-degree indicates that the number of fans of users has a relatively small impact on the information received by users. As the process of false public opinion information propagation proceeds, the S-state nodes that are finally left in the network have both small in-degree and out-degree, which corresponds to the fact that user groups or "zombie accounts" with fewer followers and fans in online social networks will not know about the public opinion topic.

[0381] Fig.13 is the change process of the degree distribution of H-state users. It can be seen that the number of H-state users in the Twitter social network is 0 at the initial state of the spread, and then begins to increase rapidly and reaches a peak at around t=70, and then gradually decreases until it basically dissipates at t=560. Its growth trend in the entire process of the spread of false public opinion information on the network is similar to the trend of the reduction of H-state users. Fig.13 Different colors in the figure represent the number of users in the H state. Fig.10 The proportion of H-state users with low entry degree is small, but their absolute number is large. Fig.13 In the figure, the expression is in red. Similarly, even when the spread of false public opinion information on Twitter social network is nearing its end, most of the remaining H states in the network are users with a small in-degree. This once again confirms that the spread of information on Twitter social network is affected by the in-degree of users, and users with a larger in-degree are more likely to spread false public opinion information.

[0382] The change process of user degree distribution in state I is as follows: Fig.14 As shown in Figure 2, during the dissemination of false public opinion information on Twitter, there are initial dissemination node users in the I state at the beginning, but since their number only accounts for 0.001 of the total number of Twitter users, Fig.14 (a) can be roughly ignored. During the propagation process, the number of users in state I gradually increases and reaches a peak at t=200, and then decreases until it returns to 0.

[0383] Will Fig.13 and Fig.14 By comparison, we can find that the time when H-state users reach their peak value (t=70) is different from the time when I-state users reach their peak value (t=200) during the spread of false public opinion information on Twitter. The latter is relatively late. Fig.13 (b) higher than Fig.14 (c) The peak value of I-state users, among which the number of H-state users with low in-degree and low out-degree is greater than the number of I-state users. When t = 560, H-state users in the Twitter network have basically disappeared, but the number of I-state users is still about half of the peak value, and the reduction of their in-degree and out-degree users is relatively uniform.

[0384] at last, Fig.15 is the change in the in-degree and out-degree distribution of the R-state node. It can be seen that during the spread of false public opinion information on Twitter social network, the number of R-state users has increased steadily, and the increase in their in-degree and out-degree users is relatively even. Compared with S-state users, the number of R-state users reaches equilibrium one step later.

[0385] In this implementation, based on a data set of a real online social platform (such as Twitter), the structure of the online social network in reality is restored, and the network is used to verify the effectiveness of the propagation dynamics model. By quantitatively analyzing and evaluating the restoration effect of the propagation process, the accuracy of the restoration of the false public opinion propagation in the real network by the heterogeneous directed online social network is evaluated. The results show that the online social network false public opinion information propagation dynamics model can more comprehensively restore the scale limitations of the real network false public opinion information propagation, and provide a theoretical basis for the application of the classic infectious disease model and directed complex network theory in the field of false public opinion information propagation, and at the same time provide a foundation for the public opinion control of online social networks.

[0386] The technical solution provided by the present invention is further described in detail above through several specific implementation modes in order to highlight the advantages and benefits of the technical solution provided by the present invention. However, the several specific implementation modes described above are not intended to be used as limitations on the present invention. Any reasonable changes and improvements to the present invention, reasonable combinations of implementation modes and equivalent substitutions based on the spirit and principles of the present invention should be included in the protection scope of the present invention.< / k> < / k> < / j> < / k> < / k> < / j>

Claims

1. The modeling method of the false public opinion information propagation dynamics model is characterized by: The method comprises the following steps: A step for constructing a directed graph network of an online social network according to a follow mechanism of the online social network; The directed graph network includes multiple nodes and multiple edges connecting any two account nodes; the nodes represent independent users in the directed online social network; the edges represent the attention relationship between subscribers and subscribers in the directed online social network; the directed graph network is ;in, Represents a collection of nodes; represents a set of edges, and ; The directed graph network satisfies the following assumptions: Hypothesis 1: Based on the Barabási-Albert model, the topological structure of the directed graph network remains approximately stable during the propagation of false public opinion information in social networks; Assumption 2: There are no self-loops or repeated connections between any account nodes in the directed graph network; Steps for constructing a propagation dynamics model: The steps for constructing a propagation dynamics model include the following steps: Steps for defining propagation rules for the spread of false public opinion information; Steps for obtaining the relationship between the probability of a node user unknown becoming a propagator and edge characteristics in an online social network: Calculate the influence of neighbor nodes on the current node and the steps of the dynamic equations for the propagation of false public opinion information in online social networks.

2. The modeling method of the false public opinion information propagation dynamics model according to claim 1 is characterized in that: The step of constructing a directed graph network of an online social network according to the attention mechanism of the online social network comprises the following steps: Steps for node classification: The account nodes in the directed graph network are divided into four types of nodes: unknown, hesitant, propagator, and immune; The unknown persons are online social network users who have not been exposed to false public opinion information so far and are still unaware of the public opinion topic; The hesitant ones are online social network users who have learned about false public opinion information but have not yet decided whether to express their opinions or views; The disseminator is an online social network user who spreads false public opinion information, opinions or views, represented by I; The immune people are online social network users who know the false public opinion information but eventually lose interest in it and stop spreading it; Steps to simplify a directed graph network: Let S represent the proportion of unknown people in the whole group, let H represent the proportion of hesitant people in the whole group, let I represent the proportion of spreaders in the whole group, let R represent the proportion of immune people in the whole group, and satisfy the following formula: ； At time t = 0, the group size of communicators who initiate discussions and spread false public opinion information in social networks is assumed to be ignored: ； The initial value of the node state ratio of the directed online social network public opinion information dissemination system is approximately: ； The step used for the degree distribution.

3. The modeling method of the false public opinion information propagation dynamics model according to claim 2 is characterized in that: The steps for degree distribution include: The degree of a node in a network refers to the number of edges connected to this node; In a directed graph network, the in-degree of a node refers to the number of edges entering the node, which indicates the number of users followed by other users; the out-degree of a node refers to the number of edges departing from the node, which indicates the number of users followed by the user; For any category : make Indicates that state A has in-degree at time t and out-degree The percentage of online social network users; make It means that an arbitrary node has The probability of degree is expressed as follows: ； ； The joint probability distribution of the probability cost of randomly selecting a node with a specific in-degree and out-degree Describe and express it as a probability generating function: In the formula, x and y Placeholder for constructing probability generating functions; Average in-degree of directed graph networks and average out-degree It is expressed by the following formulas and symbols: ； ； ； In the formula, m is the total number of in-degrees in the online social network, which is the same as the total number of out-degrees; v is the total number of nodes in the online social network.

4. The modeling method of the false public opinion information propagation dynamics model according to claim 3 is characterized in that: The steps for defining the propagation rules for the propagation of false public opinion information include: The propagation rules are as follows: Rule 1: Make another Type Node Online social network users receive false public opinion information from one of their disseminator neighbors, which is a Poisson process with a constant rate r; When the When a user at the type node is exposed to a false public opinion topic, the user's state will immediately change to a hesitant state; Rule 2: Type node users need to decide whether to spread false public opinion information; Among them, the probability of choosing to spread information related to false public opinion topics is , the user's status will be changed to communicator; The probability of remaining silent and ignoring the public opinion information is , the user's status will be changed to immune; Rule 3: One Type node users may completely lose interest in the public opinion information and The rate at which they become immune.

5. The modeling method of the false public opinion information propagation dynamics model according to claim 4 is characterized in that: The step of obtaining the relationship between the probability of an unknown node user becoming a communicator and the edge characteristics in the online social network comprises: make It represents the probability that the directed edge whose successor node state is S at time t does not transmit public opinion information, which is the percentage of S-state nodes with in-degree 1 at time t; Will Further classified by its predecessor node status: ; The relationship between these related variables is expressed by the following formula: ； In the formula, Satisfy at time t Define the proportion of directed edges whose predecessor node state is S; Satisfy at time t Define the proportion of directed edges whose predecessor node state is H; Satisfy at time t Define the proportion of directed edges whose predecessor node state is I; Satisfy at time t Define the proportion of directed edges whose predecessor node status is R; The change over time can be expressed as: ; The conditional probability that the state of the predecessor node of any directed edge is I and the state of the successor node is S can be expressed as: ： One with The expected number of directed edges of a node in state S with degree is : make Indicates that the state S is still maintained at time t The ratio of degree nodes satisfies the following differential equation: ； To earn points: ； In the formula, is the constant obtained by integration. And then get: The constant by Initial conditions influence: When false public opinion information begins to spread: → 1, we can see → 1: The analytical solution is as follows: ： By the above From the definition of: ： have The expression of the proportion of S-state nodes is: 。 6. The theme name according to claim 5, characterized in that: The step of calculating the influence of neighboring nodes on the current node and the dynamic equations of false public opinion information propagation in online social networks includes: For an arbitrary target node, let It means that the degree of any in-degree neighbor node is The probability of , which is also expressed as the degree of the predecessor node of any directed edge with the target node as the successor node is probability; In a directed graph network, the degree is The total number of nodes is , the total number of directed edges from these nodes is ; Any node has a degree of Probability of the predecessor node It is expressed as: ； It is derived from the following equation: ； The above equation can be further deduced as follows: ； all The ratio of the total out-degree of the type node to the total out-degree of all nodes in the network; and As shown below: ； ； According to the above propagation rules, The dynamic change of is determined by two parts: the proportion of H-state nodes that spread false public opinion topic information and the negative term. A portion of the state nodes become immune state R and (b) successfully infect an S state node; Satisfy the following formula: The dynamic equations for the spread of false public opinion information on online social networks are as follows: ； ； ； ； ； ； ； ； 。 7. A modeling device for a dynamic model of false public opinion information propagation, characterized in that: The device comprises the following modules: A module for constructing a directed graph network of an online social network based on the attention mechanism of the online social network; The directed graph network includes multiple nodes and multiple edges connecting any two account nodes; the nodes represent independent users in the directed online social network; the edges represent the attention relationship between subscribers and subscribers in the directed online social network; the directed graph network is ;in, Represents a collection of nodes; represents a set of edges, and ; The directed graph network satisfies the following assumptions: Hypothesis 1: Based on the Barabási-Albert model, the topological structure of the directed graph network remains approximately stable during the propagation of false public opinion information in social networks; Assumption 2: There are no self-loops or repeated connections between any account nodes in the directed graph network; Modules for constructing a propagation dynamics model: The steps for constructing a propagation dynamics model include the following modules: A module for defining propagation rules for the spread of false public opinion information; Module for obtaining the relationship between the probability of unknown node users becoming propagators and edge characteristics in online social networks: A module that calculates the influence of neighboring nodes on the current node and the dynamic equations for the propagation of false public opinion information in online social networks.

8. A computer device comprising: A processor and a memory, characterized in that the memory is used to store executable instructions of the processor, and the processor is configured to execute the modeling method of the false public opinion information propagation dynamics model described in any one of claims 1-6 by executing the executable instructions.

9. A computer storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is run, the modeling method of the false public opinion information propagation dynamics model described in any one of claims 1 to 6 is executed.

10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the steps of the method for modeling the false public opinion information propagation dynamics model described in any one of claims 1 to 6 are implemented.

Images