A rumor propagation dynamics prediction method based on rumor and resistance behavior differentiation mechanism

Abstract

This invention relates to a method for predicting the dynamics of rumor propagation based on the differentiation mechanism between rumor-mongering and resistance behaviors, belonging to the field of network propagation and control. The method includes: 1) classifying users into six mutually exclusive states according to actual application scenarios and proposing basic assumptions; 2) establishing a set of dynamic equations for the evolution of user density over time based on state transition mechanisms; 3) further calculating a basic propagation threshold and using it as a propagation discrimination index to assess the risk of continued rumor spread and its attenuation trend; 4) constructing a node state evolution algorithm based on real network datasets to verify the model's propagation patterns and predict trends; and 5) using real rumor event data for parameter fitting and prediction verification to obtain subsequent propagation evolution trends. This invention can more meticulously depict the incubation process, propagation differentiation, and resistance differences after users encounter rumors, providing technical support for risk assessment, propagation trend prediction, and intervention strategy formulation for rumor propagation on online social networks.

Description

Technical Field

[0001] This invention belongs to the field of network propagation and control, and relates to a method for predicting the dynamics of rumor propagation based on the differentiation mechanism of rumor-mongering and resistance behavior. Background Technology

[0002] With the rapid development of online social networks, users can quickly participate in information dissemination through forwarding, commenting, liking, and following, making information spread rapidly, wide-ranging, and impactful. At the same time, unverified information and false content can easily spread rapidly through user networks, forming online rumors. Especially during public emergencies, social hotspots, or public safety incidents, the spread of rumors can lead to public misunderstanding, panic, and public opinion risks. Therefore, modeling, analyzing, and predicting the rumor spread process on online social networks is of great significance.

[0003] Existing research on rumor propagation largely draws on infectious disease dynamics models, describing the rumor spread process by setting different user states and establishing state transition equations. While this approach can characterize the general patterns of rumor propagation to some extent, it still suffers from insufficiently detailed state classification. For example, some models do not adequately consider the latent stage after users are exposed to rumors but have not yet exhibited overt propagation behavior, making it difficult to reflect the transition from information exposure to manifested behavior. Furthermore, existing models typically treat rumor-spreading users or those resisting rumors as relatively singular groups, failing to adequately distinguish between different user behavior types such as normal rumor-spreading, malicious rumor-spreading, active resistance, and silent resistance. This limits the models' ability to characterize the behavioral differentiation features in real-world propagation scenarios.

[0004] Furthermore, existing models primarily remain at the level of theoretical analysis or idealized numerical simulations, lacking sufficient integration with real online social network structures and real-world rumor event data. Real-world networks exhibit complex user connection relationships and uneven node degree distribution; different network topologies significantly influence the propagation path, speed, and ultimate reach of rumors. Analysis based solely on uniform mixing assumptions or simplified network structures makes it difficult to verify the model's applicability in real-world network environments. Simultaneously, real-world rumor events are characterized by varying durations, significant differences in propagation scale, and inconsistent levels of anti-rumor participation. Without parameter fitting and predictive validation based on real-world event data, models cannot be reliably applied to predict subsequent propagation trends.

[0005] Therefore, there is a need for a modeling and prediction method for the spread of rumors on online social networks. This method should be able to refine user state segmentation and construct dynamic equations, verify the propagation patterns of the model using real network datasets, and perform parameter fitting and prediction verification using real rumor event data. This would improve the model's ability to characterize and predict the actual spread of rumors on online social networks, and provide support for rumor risk assessment and intervention strategy formulation. Summary of the Invention

[0006] In view of this, the purpose of this invention is to provide a method for predicting the dynamics of rumor propagation based on the differentiation mechanism between rumor-mongering and resistance behavior.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] A method for predicting the dynamics of rumor propagation based on the differentiation mechanism between rumor-mongering and boycott behavior includes the following steps:

[0009] Step 1: Based on users' contact status, dissemination behavior status, and resistance behavior status regarding the target rumor event, divide users in online social networks into six mutually exclusive states and make basic assumptions;

[0010] Step 2: Based on the transformation mechanism between the six mutually exclusive states, construct a set of dynamic equations to describe the change of user density over time in each state;

[0011] Step 3: Calculate the basic propagation threshold corresponding to the target rumor event based on the aforementioned dynamic equations, and determine the conditions for the continued propagation or disappearance of the target rumor event in online social networks based on the basic propagation threshold.

[0012] Step 4: Construct a real network state evolution algorithm based on a real network dataset, and perform model verification and propagation trend prediction on the propagation evolution law of the dynamic equations under different initial conditions;

[0013] Step 5: Based on the real rumor event dataset, the duration of the target rumor event is divided into a fitting phase and a verification phase; in the fitting phase, the model parameters in the dynamic equation system are fitted, and the fitted model parameters are used to make predictions and verifications in the verification phase to obtain the propagation and evolution trend of the target rumor event in the subsequent time period.

[0014] Optionally, in step one, based on the application scenario, users in the online social network are divided into six mutually exclusive states, specifically including:

[0015] Unknown users (U): These are users who have not yet received information related to the target rumor event. These users are potential users who may be affected by the subsequent spread of rumors.

[0016] Latent users (L): These are users who have been exposed to information related to the target rumor event but have not yet engaged in overt dissemination or resistance. These users are in a transitional phase between receiving information and manifesting behavior, and can be transformed into normal rumor-spreading users or malicious rumor-spreading users under subsequent information stimulation, repeated exposure, or social interaction.

[0017] Normal rumor-spreading users (BS): These are users who accept or believe in the target rumor and participate in its dissemination. Their dissemination behavior mainly manifests as general participation behaviors such as forwarding, commenting, discussing, or following up on the event. Their dissemination methods are relatively restrained and usually do not aim to incite, manipulate, or maliciously spread rumors.

[0018] Malicious rumor-spreading users (BI): These are users who participate in the spread of a target rumor event and whose spreading behavior is highly aggressive or manipulative. Such users' behavior includes at least one of the following: exaggerating facts, inciting emotions, guiding public opinion, or organizing the spread. Their spreading intensity and negative impact are higher than those of normal rumor-spreading users.

[0019] Resistant Users (RI): Users who take proactive measures to resist a target rumor. These users influence other users and inhibit the spread of the target rumor by posting debunking content, posting error correction information, forwarding authoritative clarification content, or outputting reverse information.

[0020] Resisting Inactive Users (RS): This refers to users who choose to stop participating in the dissemination of the target rumor after verifying the target rumor or accessing debunking information. These users will no longer forward, comment on, discuss, or publish information related to the target rumor and will maintain a state of resistance to the target rumor by remaining silent.

[0021] Furthermore, by introducing relevant model parameters, we make assumptions and characterize the user state transformation mechanism in the process of rumor propagation:

[0022] (i) External users at rate Once you enter an online social network, users who are already online, regardless of their current status, may experience an offline rate. Leave the system. Once logged out, the user will no longer participate in the spread, transformation, or resistance of the targeted rumor event.

[0023] (ii) When unknown users encounter content related to rumors on online social networks, they are affected by both normal rumor-spreading users and malicious rumor-spreading users, and the impact is expressed by parameters. , It represents the probability of a state changing from an unknown state to a latent state.

[0024] (iii) After continuously receiving relevant information, participating in discussions, or being stimulated by repeated information, latent users will gradually form an attitude judgment towards the target rumor event, and will then express their opinions accordingly. , The probability of being converted into a normal rumor-spreading user or a malicious rumor-spreading user.

[0025] (iv) Normal users spreading rumors may be influenced by factors such as group emotions, herd mentality, adversarial interactions, or information distortion during the dissemination process, which may further radicalize their dissemination behavior and lead to... The probability of this turning into malicious rumor-spreading users.

[0026] (v) As facts become clearer, authoritative clarifications spread, or platform interventions take effect, normal users spreading rumors can... The probability of this is transformed into resistance against users with no behavior, and with The probability of this translates into resistance against users who engage in such behavior; users who maliciously spread rumors can, under corrective, restrictive, or interventional measures, [achieve their desired outcome]. The probability of this translates into resistance from users who exhibit certain behaviors.

[0027] (vi) Resisting users whose active voice may decrease over time after completing the dissemination of debunking, correcting, or clarifying information, and whose... The probability of this translates into resistance against users who exhibit no behavior.

[0028] (vii) Resisting users who have ceased participating in the posting, forwarding, commenting or discussing of information related to the target rumor event; these users no longer spread rumors, nor do they actively provide debunking information, and are in a state of silent resistance.

[0029] Optionally, in step two, the density of users in each state can be determined by... , , , , as well as This indicates that, for ease of expression, it is abbreviated as [insert abbreviation here]. , , , , as well as .also, This represents the average degree of the network. Based on the definitions of state variables and assumptions about state transition mechanisms, the following set of mean-field dynamic equations is established to describe the evolution of the state density of various user types over time:

[0030]

[0031] Optionally, in step three, the next-generation matrix method is used to select the infection variable. The equations corresponding to the propagation-related state variables in the dynamic equation system are split into new propagation generation terms and state transition terms, i.e. Next, by calculating the equilibrium point in the absence of rumors... From the Jacobian matrices of F(X) and V(X), we can obtain:

[0032] ,

[0033] Let A = B= C= Then the matrix We can obtain:

[0034]

[0035] Then the spectral radius of matrix K is equal to the fundamental regeneration number of the dynamical system:

[0036]

[0037] when When the target rumor event is determined to not meet the conditions for sustained spread on online social networks, the propagation system tends towards a rumor-free equilibrium state; when When a target rumor event is determined to meet the conditions for sustained spread on online social networks, the dissemination system is deemed to have a risk of continued rumor spread.

[0038] Optionally, in step four, a real network state evolution algorithm is constructed and executed on real network datasets such as Facebook, Twitter, and P2P. The results of the change in the number of user nodes obtained by the real network state evolution algorithm are compared with the numerical solution results of the dynamic equations under different initial population ratios and different system parameters to verify the ability of the dynamic equations to characterize the propagation and evolution of real networks.

[0039] Input: Set the initial network Output: Number of nodes 1. Initialize nodes: Randomly assign the initial number of nodes in different states. 2. Set the iteration count to K. 3. For t = 1 to K, with a step size of 1. 4. If node i is in state U at time t and does not leave the network at time t+1, then traverse the neighboring nodes. Under the influence of nodes BS and BI, i... The probability becomes L node ( , (This represents the number of neighbors of node U in BS or BI states at time t) 5. else Node i's state remains unchanged. 6. else if node i is in state L at time t and does not leave the network at time t+1, then 7. Node i... or 8. If node i's probability changes to BS or BI, then 9. If node i is in BS state at time t and does not leave the network at time t+1, then 10. Node i... , or The probability of node 11 changes to BI, RS, or RI. Otherwise, the state of node i remains unchanged. 12. else if node i is in state BI at time t and does not leave the network at time t+1, then 13. Node i... The probability of node i becoming RI node 14. else if node i is in RI state at time t and does not leave the network at time t+1, then 15. Node i becomes RI node 14. The probability of node i changes to RS. 16. else, the state of node i remains unchanged. 17. end if 18. end for 19. Calculate Count the number of nodes and return the result.

[0040] Optionally, in step five, based on real rumor events on the Twitter platform, the duration of the target rumor event is divided into a fitting stage and a verification stage. In the fitting stage, the model parameters in the dynamic equation system are fitted, and the fitted model parameters are used to make predictions and verifications in the verification stage to obtain the propagation and evolution trend of the target rumor event in the subsequent time period.

[0041] The beneficial effects of this invention are as follows: This invention proposes a method for predicting the dynamics of rumor propagation based on the differentiation mechanism of rumor-mongering and resistance behavior, including: 1) By dividing users into six mutually exclusive states, it more meticulously depicts the evolutionary process of users from encountering rumors, latent judgment, differentiated propagation to differentiated resistance; 2) By constructing a dynamic equation describing the change of user density over time in various states, it calculates the basic propagation threshold corresponding to the target rumor event, which can determine the conditions for the continued propagation or disappearance of rumors in online social networks, providing a basis for risk assessment of rumor spread; 3) By constructing a real network state evolution algorithm based on real network datasets and comparing the real network evolution results with the numerical results of the dynamic equations, it enhances the applicability and credibility of the model in real network topology environments; 4) By using real rumor event datasets for parameter fitting and prediction verification, it can obtain the propagation evolution trend of the target rumor event in subsequent time periods, improve the accuracy of rumor propagation trend prediction, and provide support for the formulation of online social network rumor governance and intervention strategies. Attached Figure Description

[0042] To make the objectives, technical solutions, and advantages of the present invention clearer, the preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, wherein:

[0043] Figure 1 The present invention provides a detailed implementation flowchart;

[0044] Figure 2 This is an application scenario diagram of the present invention;

[0045] Figure 3 This is the state transition diagram corresponding to the rumor propagation model of this invention. Detailed Implementation

[0046] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0047] The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual pictures. They should not be construed as limiting the invention. To better illustrate the embodiments of the invention, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0048] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," "front," and "rear" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present invention. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0049] See Figures 1-3 This invention provides a method for predicting the dynamics of rumor propagation based on the differentiation mechanism between rumor-mongering and boycott behavior. Figure 1 The flowchart is for the specific implementation. Figure 2 This is a diagram illustrating an application scenario of the present invention. Figure 3 This is the state transition diagram corresponding to the rumor propagation model. The following explanation, with reference to the attached diagram, includes the following steps:

[0050] Optionally, in step one, based on the application scenario, users in the online social network are divided into six mutually exclusive states, specifically including:

[0051] Unknown users (U): These are users who have not yet received information related to the target rumor event. These users are potential users who may be affected by the subsequent spread of rumors.

[0052] Latent users (L): These are users who have been exposed to information related to the target rumor event but have not yet engaged in overt dissemination or resistance. These users are in a transitional phase between receiving information and manifesting behavior, and can be transformed into normal rumor-spreading users or malicious rumor-spreading users under subsequent information stimulation, repeated exposure, or social interaction.

[0053] Normal rumor-spreading users (BS): These are users who accept or believe in the target rumor and participate in its dissemination. Their dissemination behavior mainly manifests as general participation behaviors such as forwarding, commenting, discussing, or following up on the event. Their dissemination methods are relatively restrained and usually do not aim to incite, manipulate, or maliciously spread rumors.

[0054] Malicious rumor-spreading users (BI): These are users who participate in the spread of a target rumor event and whose spreading behavior is highly aggressive or manipulative. Such users' behavior includes at least one of the following: exaggerating facts, inciting emotions, guiding public opinion, or organizing the spread. Their spreading intensity and negative impact are higher than those of normal rumor-spreading users.

[0055] Resistant Users (RI): Users who take proactive measures to resist a target rumor. These users influence other users and inhibit the spread of the target rumor by posting debunking content, posting error correction information, forwarding authoritative clarification content, or outputting reverse information.

[0056] Resisting Inactive Users (RS): This refers to users who choose to stop participating in the dissemination of the target rumor after verifying the target rumor or accessing debunking information. These users will no longer forward, comment on, discuss, or publish information related to the target rumor and will maintain a state of resistance to the target rumor by remaining silent.

[0057] Furthermore, by introducing relevant model parameters, we make assumptions and characterize the user state transformation mechanism in the process of rumor propagation:

[0058] (i) External users at rate Once you enter an online social network, users who are already online, regardless of their current status, may experience an offline rate. Leave the system. Once logged out, the user will no longer participate in the spread, transformation, or resistance of the targeted rumor event.

[0059] (ii) When unknown users encounter content related to rumors on online social networks, they are affected by both normal rumor-spreading users and malicious rumor-spreading users, and the impact is expressed by parameters. , It represents the probability of a state changing from an unknown state to a latent state.

[0060] (iii) After continuously receiving relevant information, participating in discussions, or being stimulated by repeated information, latent users will gradually form an attitude judgment towards the target rumor event, and will then express their opinions accordingly. , The probability of being converted into a normal rumor-spreading user or a malicious rumor-spreading user.

[0061] (iv) Normal users spreading rumors may be influenced by factors such as group emotions, herd mentality, adversarial interactions, or information distortion during the dissemination process, which may further radicalize their dissemination behavior and lead to... The probability of this turning into malicious rumor-spreading users.

[0062] (v) As facts become clearer, authoritative clarifications spread, or platform interventions take effect, normal users spreading rumors can... The probability of this is transformed into resistance against users with no behavior, and with The probability of this translates into resistance against users who engage in such behavior; users who maliciously spread rumors can, under corrective, restrictive, or interventional measures, [achieve their desired outcome]. The probability of this translates into resistance from users who exhibit certain behaviors.

[0063] (vi) Resisting users whose active voice may decrease over time after completing the dissemination of debunking, correcting, or clarifying information, and whose... The probability of this translates into resistance against users who exhibit no behavior.

[0064] (vii) Resisting users who have ceased participating in the posting, forwarding, commenting or discussing of information related to the target rumor event; these users no longer spread rumors, nor do they actively provide debunking information, and are in a state of silent resistance.

[0065] Optionally, in step two, the density of users in each state can be determined by... , , , , as well as This indicates that, for ease of expression, it is abbreviated as [insert abbreviation here]. , , , , as well as .also, This represents the average degree of the network. Based on the definitions of state variables and assumptions about state transition mechanisms, the following set of mean-field dynamic equations is established to describe the evolution of the state density of various user types over time:

[0066]

[0067] Optionally, in step three, the next-generation matrix method is used to select the infection variable. The equations corresponding to the propagation-related state variables in the dynamic equation system are split into new propagation generation terms and state transition terms, i.e. Next, by calculating the equilibrium point in the absence of rumors... From the Jacobian matrices of F(X) and V(X), we can obtain:

[0068] ,

[0069] Let A = B= C= Then the matrix We can obtain:

[0070]

[0071] Then the spectral radius of matrix K is equal to the fundamental regeneration number of the dynamical system:

[0072]

[0073] when When the target rumor event is determined to not meet the conditions for sustained spread on online social networks, the propagation system tends towards a rumor-free equilibrium state; when When a target rumor event is determined to meet the conditions for sustained spread on online social networks, the dissemination system is deemed to have a risk of continued rumor spread.

[0074] Optionally, in step four, a real network state evolution algorithm is constructed and executed on real network datasets such as Facebook, Twitter, and P2P. The results of the change in the number of user nodes obtained by the real network state evolution algorithm are compared with the numerical solution results of the dynamic equations under different initial population ratios and different system parameters to verify the ability of the dynamic equations to characterize the propagation and evolution of real networks.

[0075] Input: Set the initial network Output: Number of nodes 1. Initialize nodes: Randomly assign the initial number of nodes in different states. 2. Set the iteration count to K. 3. For t = 1 to K, with a step size of 1. 4. If node i is in state U at time t and does not leave the network at time t+1, then traverse the neighboring nodes. Under the influence of nodes BS and BI, i... The probability becomes L node ( , (This represents the number of neighbors of node U in BS or BI states at time t) 5. else Node i's state remains unchanged. 6. else if node i is in state L at time t and does not leave the network at time t+1, then 7. Node i... or 8. If node i's probability changes to BS or BI, then 9. If node i is in BS state at time t and does not leave the network at time t+1, then 10. Node i... , or The probability of node 11 changes to BI, RS, or RI. Otherwise, the state of node i remains unchanged. 12. else if node i is in state BI at time t and does not leave the network at time t+1, then 13. Node i... The probability of node i becoming RI node 14. else if node i is in RI state at time t and does not leave the network at time t+1, then 15. Node i becomes RI node 14. The probability of node i changes to RS. 16. else, the state of node i remains unchanged. 17. end if 18. end for 19. Calculate Count the number of nodes and return the result.

[0076] Optionally, in step five, based on real rumor events on the Twitter platform, the duration of the target rumor event is divided into a fitting stage and a verification stage. In the fitting stage, the model parameters in the dynamic equation system are fitted, and the fitted model parameters are used to make predictions and verifications in the verification stage to obtain the propagation and evolution trend of the target rumor event in the subsequent time period.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for predicting the dynamics of rumor propagation based on the differentiation mechanism between rumor-mongering and boycott behavior, characterized in that: The method includes the following steps: Step 1: Based on users' contact status, dissemination behavior status, and resistance behavior status regarding the target rumor event, divide users in online social networks into six mutually exclusive states and make basic assumptions; Step 2: Based on the transformation mechanism between the six mutually exclusive states, construct a set of dynamic equations to describe the change of user density over time in each state; Step 3: Calculate the basic propagation threshold corresponding to the target rumor event based on the aforementioned dynamic equations, and determine the conditions for the continued propagation or disappearance of the target rumor event in online social networks based on the basic propagation threshold. Step 4: Construct a real network state evolution algorithm based on a real network dataset, and perform model verification and propagation trend prediction on the propagation evolution law of the dynamic equations under different initial conditions; Step 5: Based on the real rumor event dataset, the duration of the target rumor event is divided into a fitting phase and a verification phase; in the fitting phase, the model parameters in the dynamic equation system are fitted, and the fitted model parameters are used to make predictions and verifications in the verification phase to obtain the propagation and evolution trend of the target rumor event in the subsequent time period.

2. The rumor propagation dynamics prediction method based on the differentiation mechanism of rumor-mongering and resistance behavior as described in claim 1, characterized in that: The specific process of step one includes: dividing users in online social networks into six mutually exclusive states based on the application scenario, specifically including: Unknown users (U): These are users who have not yet received information related to the target rumor event. These users are potential users who may be affected by the subsequent spread of rumors. Latent users (L): These are users who have been exposed to information related to the target rumor event but have not yet engaged in overt dissemination or resistance. These users are in a transitional phase between receiving information and manifesting behavior, and can be transformed into normal rumor-spreading users or malicious rumor-spreading users under subsequent information stimulation, repeated exposure, or social interaction. Normal rumor-spreading users (BS): These are users who accept or believe in the target rumor and participate in its dissemination. Their dissemination behavior mainly manifests as general participation behaviors such as forwarding, commenting, discussing, or following up on the event. Their dissemination methods are relatively restrained and usually do not aim to incite, manipulate, or maliciously spread rumors. Malicious rumor-spreading users (BI): These are users who participate in the spread of a target rumor event and whose spreading behavior is highly aggressive or manipulative. Such users' behavior includes at least one of the following: exaggerating facts, inciting emotions, guiding public opinion, or organizing the spread. Their spreading intensity and negative impact are higher than those of normal rumor-spreading users. Resistant Users (RI): Users who take proactive measures to resist a target rumor. These users influence other users and inhibit the spread of the target rumor by posting debunking content, posting error correction information, forwarding authoritative clarification content, or outputting reverse information. Resisting Inactive Users (RS): This refers to users who choose to stop participating in the dissemination of the target rumor after verifying the target rumor or accessing debunking information. These users will no longer forward, comment on, discuss, or publish information related to the target rumor and will maintain a state of resistance to the target rumor by remaining silent. Furthermore, by introducing relevant model parameters, we make assumptions and characterize the user state transformation mechanism in the process of rumor propagation: (i) External users at rate Once you enter an online social network, users who are already online, regardless of their current status, may experience an offline rate. Leave the system. Once logged out, the user will no longer participate in the spread, transformation, or resistance of the targeted rumor event. (ii) When unknown users encounter content related to rumors on online social networks, they are affected by both normal rumor-spreading users and malicious rumor-spreading users, and the impact is expressed by parameters. , It represents the probability of a state changing from an unknown state to a latent state. (iii) After continuously receiving relevant information, participating in discussions, or being stimulated by repeated information, latent users will gradually form an attitude judgment towards the target rumor event, and will then express their opinions accordingly. , The probability of being converted into a normal rumor-spreading user or a malicious rumor-spreading user. (iv) Normal users spreading rumors may be influenced by factors such as group emotions, herd mentality, adversarial interactions, or information distortion during the dissemination process, which may further radicalize their dissemination behavior and lead to... The probability of this turning into malicious rumor-spreading users. (v) As facts become clearer, authoritative clarifications spread, or platform interventions take effect, normal users spreading rumors can... The probability of this is transformed into resistance against users with no behavior, and with The probability of this translates into resistance against users who engage in such behavior; users who maliciously spread rumors can, under corrective, restrictive, or interventional measures, [achieve their desired outcome]. The probability of this translates into resistance from users who exhibit certain behaviors. (vi) Resisting users whose active voice may decrease over time after completing the dissemination of debunking, correcting, or clarifying information, and whose... The probability of this translates into resistance against users who exhibit no behavior. (vii) Resisting users who have ceased participating in the posting, forwarding, commenting or discussing of information related to the target rumor event; these users no longer spread rumors, nor do they actively provide debunking information, and are in a state of silent resistance.

3. The rumor propagation dynamics prediction method based on the differentiation mechanism of rumor-mongering and resistance behavior as described in claim 2, characterized in that, The specific process of step two includes: the density of users in each state can be determined by... , , , , as well as This indicates that, for ease of expression, it is abbreviated as [insert abbreviation here]. , , , , as well as .also, This represents the average degree of the network. Based on the definitions of state variables and assumptions about state transition mechanisms, the following set of mean-field dynamic equations is established to describe the evolution of the state density of various user types over time:

4. The rumor propagation dynamics prediction method based on the differentiation mechanism of rumor-mongering and resistance behavior as described in claim 3, characterized in that, The specific process of step three includes: selecting infection variables using the next-generation matrix method. The equations corresponding to the propagation-related state variables in the dynamic equation system are split into new propagation generation terms and state transition terms, i.e. Next, by calculating the equilibrium point in the absence of rumors... From the Jacobian matrices of F(X) and V(X), we can obtain: , Let A = B= C= Then the matrix We can obtain: Then the spectral radius of matrix K is equal to the fundamental regeneration number of the dynamical system: when When the target rumor event is determined to not meet the conditions for sustained spread on online social networks, the propagation system tends towards a rumor-free equilibrium state; when When a target rumor event is determined to meet the conditions for sustained spread on online social networks, the dissemination system is deemed to have a risk of continued rumor spread.

5. The rumor propagation dynamics prediction method based on the differentiation mechanism of rumor-mongering and resistance behavior as described in claim 1, characterized in that, The specific process of step four includes: constructing a real network state evolution algorithm, and comparing the changes in the number of user nodes obtained by the real network state evolution algorithm with the numerical solution results of the dynamic equations under different initial conditions based on real network Facebook, Twitter and P2P datasets, in order to verify the ability of the dynamic equations to characterize the propagation and evolution law of real networks.

6. The rumor propagation dynamics prediction method based on the differentiation mechanism of rumor-mongering and resistance behavior according to claim 1, characterized in that, The specific process of step five includes, based on real rumor events on the Twitter platform, dividing the duration of the target rumor event into a fitting stage and a verification stage. In the fitting stage, the model parameters in the dynamic equation system are fitted, and the fitted model parameters are used to make predictions and verifications in the verification stage to obtain the propagation and evolution trend of the target rumor event in the subsequent time period.

Images