CN115132369A

CN115132369A - Information propagation analysis method and system based on social media mimicry environment modeling

Info

Publication number: CN115132369A
Application number: CN202210645833.8A
Authority: CN
Inventors: 殷复莲; 唐鑫夷; 梁彤宇; 胡芷溦; 陈卓; 吴杰玲; 王锦霞; 潘妍妍; 佘雨薇; 夏欣雨
Original assignee: Communication University of China
Current assignee: Communication University of China
Priority date: 2022-06-09
Filing date: 2022-06-09
Publication date: 2022-09-30
Anticipated expiration: 2042-06-09
Also published as: CN115132369B

Abstract

The invention provides an information propagation analysis method and system based on social media mimicry environment modeling, which analyze two propagation links of social media platform information propagation by constructing an SFI-PE model under the visual threshold of an information ecosystem and an M-SFI-PE model under the intervention action of a third party considering factors outside the system, comprehensively research the interaction and dynamic influence between people and between environments in the social media information ecosystem, and summarize the general rule of the network information propagation ecosystem more suitable for the current new media environment by combining the parameter sensitivity analysis of information propagation indexes. Aiming at the promotion propagation of positive events and the inhibition propagation experiment of negative events, the effective method for information propagation intervention of third party intervention is obtained, and the fact that the user who is easily influenced contacts with the information through social relations is determinedAverage forwarding probability p generated after arrival of information _F Is the key cut-in point and the optimal direction of the third-party intervention.

Description

Information propagation analysis method and system based on social media mimicry environment modeling

Technical Field

The invention relates to the technical field of information propagation analysis based on model construction, in particular to an information propagation analysis method and system based on social media mimicry environment modeling.

Background

In the common sense definition, "transmission" is often considered as the transmission process of information between people, so that analyzing and interpreting information transmission by taking the intrinsic characteristics of things as an incision point is a common research path, but the common research path generally ignores the importance of the information environment in the information transmission process. The people are group organisms with social attributes, the communication interaction between people needs to be generated in a certain information space, and a community of a specific topic is formed according to different information contents, so that the behaviors of the people are influenced not only by self subjectivity but also by other individuals and the environment.

Based on this, the american scholars Davenport T H first proposed the concept of "information ecology" which introduces the natural ecosystem studied by ecology into human society. The information ecosystem refers to a unified whole formed by interaction of information, information people and information ecological environment in a certain information space due to information exchange relationship, and the unified whole forms basic elements of information ecology together. Specifically, the information ecosystem includes a process of constantly exchanging and circulating information between people, human organizations, communities and their information environments. The information ecological environment refers to other information persons, information contents, information technologies, information space-time, information systems and the like which directly affect the survival, life and development of the information persons.

In the field of mass propagation, the information ecological environment corresponds to a propagation concept 'mimicry environment' proposed by waalt lippmann of the U.S. famous politicians to a certain extent. The mimicry environment is an information environment formed by mass-propagation activities, which is not a mirror-type reproduction of the real environment, but an environment displayed to people by a mass-propagation medium through selecting, processing and restructuring object characteristic events or information. In the modern society of the new media information age, the social media platform is mainly embodied as a topic community built by relying on the internet technology, such as hot search and super talk of the Xinlang microblog in China, trend of foreign platforms Twitter and Facebook and the like. On a social media platform, a topic community of a specific popular event provides soil for the transmission and discussion of the event, relevant information of multi-element main body publishing, forwarding and commenting is collected, comprehensive sequencing of information presentation is realized by means of an algorithm mechanism of the platform, an information set which can reflect most user view attitudes or has advantage transmission power is screened out, and the specific popular event can be considered to be reflected on the specific platform in a centralized mode.

Currently, the most commonly used research approach for information dissemination in social networks is the model of infectious diseases. The infectious disease model is provided for researching the problems of the transmission speed, the spatial range, the transmission path, the kinetic mechanism and the like of the infectious disease so as to guide the effective prevention and control of the infectious disease. The classical infectious disease models SI (safe-Infected) and SIR (safe-Infected-Recovered) mainly consider the transmission path of human infectious agents. Later, researchers have further introduced modules of environments to study the mechanisms of infection of humans by environments with pathogens present.

Because the infection mechanism of infectious diseases has similarity with the human information transmission paradigm, the infectious disease model has been popularized to the field of social scientific research. In modern society, how to make effective network information propagation intervention strategies and improve the self-regulation capability of the network ecological environment becomes the most urgent and important subject to be faced for a long time in social governance. Therefore, students propose a public opinion transmission and infectious disease control model after a major emergency, divide the masses into susceptible infected groups, latent groups, infected groups and recovery groups, and study the dynamic influence of third-party intervention on different groups. However, in the network society constructed by the social media platform, information transmission is not only generated by means of the social relationship between people, and the mimicry environment of topic community can also directly influence the information transmission behavior of people, just like infectious diseases in an infectious disease model can infect people through natural environment. Therefore, a network information propagation analysis method based on a social media mimicry environment needs to be further researched based on a novel network society construction mechanism.

Disclosure of Invention

In view of the above problems, the present invention researches the information propagation dynamics mode of social media from a brand-new information ecological threshold, adopts the research idea of the system theory, and regards the social media platform as an information ecosystem formed by integrating information, information people and information environment (i.e. mimicry environment). The network information propagation rule is summarized by constructing an information propagation dynamic model to research the internal interaction and dynamic influence between people and between environments in the network information propagation process. And on the basis of a social media information ecosystem, the influence of each link of an external factor intervention system on network information transmission is comprehensively considered, an information transmission dynamic model based on a social media mimicry environment is constructed, an effective network information transmission analysis strategy is assisted to be made, and the self-regulation capability of the network ecological environment is improved.

According to one aspect of the invention, an information propagation dynamic model modeling method based on social media mimicry environment is provided, which is characterized in that, in the information propagation dynamic model,

in the development process of a horological event, information propagation generated by depending on social relations and information propagation generated by depending on a mimicry environment are two parallel propagation links which are not interfered with each other and have different information metabolism modes; information is spread in a closed and stable environment, the total number N of people in the environment is unchanged, N is divided into three groups of a susceptible state S, a forwarding state F and an immune state I, and the condition that each individual in the group is unique at any time is assumed; the meanings of the states in the information propagation dynamic model are as follows:

susceptible state S: the individual in the state is not exposed to the information, but is likely to be exposed to the information in the future and influenced by the information, so that the forwarding behavior is generated;

a forwarding state F: the individual in the state generates the forwarding behavior and has the capability of infecting the individual in the susceptible state to forward the information;

immune status I: the population in this state consists of two parts:

the individual in the forwarding state exceeds the active exposure period and no longer has the ability to affect others, thereby converting to the immune state; and (c) a second step of,

after contacting information, an individual in an easily affected state is directly converted into an immune state because the information is not subjectively interested;

defining B (t) as an effective propagation information set in a mimicry environment at time t, defining S (t), F (t) and I (t) as the number of instantaneous groups in each state at time t, and defining S (t) + F (t) + I (t) ═ N;

the differential equation of the information propagation dynamic model is as follows:

wherein ,β_F Average contact rate at which information can be contacted by social relationships for vulnerable users; beta is a beta _B Average contact rate at which information can be contacted by a sensitive user through a mimicry environment，p _F Average forwarding probability, p, generated after the vulnerable users contact the information through social relations _B Average forwarding probability, alpha, generated after the vulnerable user is exposed to the information through the mimicry environment _F To forward the average immunization rate at which the user becomes inactive during the information dissemination, α _B The average metabolic rate of the information in the mimicry environment becoming inactive in the information propagation process is shown, and gamma is the average exposure probability of the information forwarded by the forwarding user and presented in the mimicry environment through the screening of a platform algorithm mechanism.

According to another aspect of the present invention, an information propagation analysis method based on modeling of a social media mimicry environment is provided, which is used for performing information propagation analysis by using an information propagation dynamical model based on a social media mimicry environment, wherein the information propagation dynamical model based on a social media mimicry environment is an information propagation dynamical model based on a social media mimicry environment that is established by using the information propagation dynamical model modeling method based on a social media mimicry environment; the method comprises the following steps:

performing information transmission analysis by using the information transmission dynamic model based on the social media mimicry environment; and performing information transmission analysis on the basis of the information transmission dynamic model based on the social media mimicry environment by combining with a third-party intervention information transmission dynamic model based on the social media mimicry environment;

the method for carrying out information propagation analysis by using the information propagation dynamic model based on the social media mimicry environment comprises the following steps:

searching and collecting original information data of an event to be analyzed on a preset social platform based on a preset search engine, wherein the information data comprises a user forwarding text and forwarding time;

preprocessing the original information data to obtain a noiseless redundant user forwarding text and forwarding time under each piece of information;

carrying out data fitting and parameter estimation by taking the noiseless redundant user forwarding texts and the forwarding time as data driving; wherein parameters of the information propagation dynamic model based on the social media mimicry environment and the initial susceptible population are estimated by using a least square method;

and determining information propagation index data of the event to be analyzed through the information propagation dynamic model based on the social media mimicry environment based on the parameters and the initial susceptible population total number.

According to still another aspect of the present invention, there is provided an information dissemination analysis system based on social media mimicry environment modeling, comprising:

the modeling unit is used for creating an information propagation dynamic model based on the social media mimicry environment and a third-party intervention information propagation dynamic model based on the social media mimicry environment; the information transmission dynamic model based on the social media mimicry environment is an information transmission dynamic model based on the social media mimicry environment, which is established by the information transmission dynamic model modeling method based on the social media mimicry environment in

claim

1 or 2;

the information propagation analysis unit comprises a primary analysis unit and an intervention analysis unit, wherein the primary analysis unit is used for carrying out information propagation analysis by utilizing the information propagation dynamic model based on the social media mimicry environment; the intervention analysis unit is used for carrying out information transmission analysis on the basis of the information transmission dynamic model based on the social media mimicry environment by combining a third-party intervention information transmission dynamic model based on the social media mimicry environment;

wherein the primary analysis unit comprises:

the system comprises an information acquisition unit, a processing unit and a processing unit, wherein the information acquisition unit is used for searching and acquiring original information data of an event to be analyzed on a preset social platform based on a preset search engine, and the information data comprises a user forwarding text and forwarding time;

the information preprocessing unit is used for preprocessing the original information data to acquire a user forwarding text without noise redundancy and forwarding time under each piece of information;

the data fitting and parameter estimation unit is used for carrying out data fitting and parameter estimation by taking the user forwarding texts without noise redundancy and the forwarding time as data driving; estimating parameters and initial susceptible population total of the information propagation dynamic model based on the social media mimicry environment by using a least square method;

and the index data determining unit is used for determining the information propagation index data of the event to be analyzed through the information propagation dynamic model based on the social media mimicry environment based on the parameters and the initial susceptible population.

The invention constructs an SFI-PE model under the visual threshold of an information ecosystem, further considers the influence of each link of a third-party intervention system on information transmission on the basis of the SFI-PE model, constructs an M-SFI-PE model under the intervention action of a third party considering factors outside the system, and summarizes the general rule of the network information transmission ecosystem under the current new media environment by deeply analyzing two transmission links of the social media platform information transmission, namely the social relation and the mimicry environment constructed by depending on a platform mechanism, comprehensively researching the interaction and the dynamic influence between people and the environment in the social media information ecosystem and combining the parameter sensitivity analysis on information transmission indexes. Aiming at the promotion propagation of positive events and the inhibition propagation experiment of negative events, the effective method for information propagation intervention of third party intervention is obtained, and the average forwarding probability p generated after the susceptible users contact the information through the social relation is determined _F Is the key cut-in point and the optimal direction of the third-party intervention.

To the accomplishment of the foregoing and related ends, one or more aspects of the invention comprise the features hereinafter fully described and particularly pointed out in the claims. The following description and the annexed drawings set forth in detail certain illustrative aspects of the invention. These aspects are indicative, however, of but a few of the various ways in which the principles of the invention may be employed. Further, the present invention is intended to include all such aspects and their equivalents.

Drawings

Other objects and results of the present invention will become more apparent and readily appreciated as the same becomes better understood by reference to the following description and appended claims, taken in conjunction with the accompanying drawings. In the drawings:

FIGS. 1a and 1b are schematic diagrams of an SFI-PE model M-SFI-PE, respectively, according to the present invention;

FIG. 2 is a flow chart of an information dissemination analysis method based on social media mimicry environment modeling according to the present invention;

FIG. 3 is an information propagation curve for a specific positive event and a specific negative event in accordance with the present invention;

FIG. 4 shows an index F under the influence of various parameters according to the present invention _max 、C _s 、

The PRCC result of (a);

FIG. 5 is a diagram illustrating information propagation indicator fluctuation caused by a single parameter change according to the present invention;

FIG. 6 is a schematic diagram of the population evolution for a third party to facilitate the propagation of positive event information in accordance with the present invention;

fig. 7 is a schematic diagram of the population evolution of a third party suppressing the propagation of negative event information according to the present invention.

The same reference numbers in all figures indicate similar or corresponding features or functions.

Detailed Description

In the following description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of one or more embodiments. It may be evident, however, that such embodiment(s) may be practiced without these specific details. In other instances, well-known structures and devices are shown in block diagram form in order to facilitate describing one or more embodiments.

Through deep analysis of the operation logic and the algorithm mechanism of the social media platform, the information people and the information ecological environment, namely the topic community mimicry environment constructed by the social media platform internet technology, are an inseparable organic whole. Therefore, the invention can adapt to the current ecosystem of network information transmission by reconstructing the traditional infectious disease model and considering the dynamic influence of third party intervention on different crowds, crowds and environments, so as to provide the following information transmission analysis method and system based on social media mimicry environment modeling.

Specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

According to the information propagation analysis method and system based on the social media mimicry environment modeling, firstly, an information propagation dynamic model based on the social media mimicry environment needs to be constructed. Therefore, the invention firstly provides an information propagation dynamics model modeling method based on the social media mimicry environment.

The internet technology of social media platforms has determined that users can access information through two ways: 1. the friend relationship is a social relationship constructed by platform mechanisms such as vermicelli, attention and the like, and a user can directly contact the information forwarded by the friend on a personal homepage; 2. the topic community is a mimicry environment (hereinafter referred to as "mimicry environment") constructed by platform mechanisms such as "hot" and "keyword search", and a user can enter a related page of a specific topic or event to contact information of other people in the environment.

Meanwhile, the algorithm of the social media platform needs to comprehensively consider the time for information publishing, the content quality and the popularity of the information, the activity degree of an account number for publishing the information, whether a publisher authenticates the real name and the like, and screens and filters the information based on the factors and displays the information in a topic community in sequence, and the algorithm mechanism determines that the information contacted by a user through the topic community is not all information under the topic, but the information is presented to the public after being selected, processed and structured again.

Based on the current status of the social media platform mechanism, the information propagation dynamic model based on the social media mimicry environment, namely a social-forwarding-immune-environment (SFI-PE) model, shown in fig. 1 needs to be constructed. In the SFI-PE model, the information propagation generated by depending on the social relationship and the information propagation generated by depending on the mimicry environment are assumed to be two parallel propagation links which do not interfere with each other and have different information metabolism modes in the development process of the objection event. For example, although a user has lost the ability to propagate among friends, the information forwarded by the user may still contact other users in the "topic community".

All parameters involved in the SFI-PE model and the corresponding explanations are shown in Table 1 below.

TABLE 1SFI-PE MODEL PARAMETERS TABLE

For a particular network trending event of a social media platform, it is assumed that information dissemination is performed in a closed and stable space with a constant head count (N) in the space. In the present model, s (t), f (t), and i (t) are defined as the number of instantaneous groups in each state at time t, and s (t) + f (t) + i (t) ═ N. The population in N is subdivided into three states, namely, a vulnerable state (S), a forwarding state (F), and an immune state (I), and it is assumed that each individual in the population is in a unique state at any time. The meanings of the states in the SFI-PE model are as follows:

1. susceptible state (S): the individual in this state has not yet been exposed to the information, but may be exposed to and influenced by the information in the future, resulting in a forwarding action.

2. Forwarding state (F): individuals in this state produce a forwarding action with the ability to infect individuals in a vulnerable state to forward information.

3. Immune status (I): the group in the state mainly comprises two parts, namely, the individual in the forwarding state exceeds the active exposure period and is no longer capable of influencing other people, so that the individual is converted into the immune state; secondly, after contacting information, the individual in the easily affected state is directly converted into the immune state because the information is not subjectively interested.

Meanwhile, B (t) is defined in the SFI-PE model to be an effective propagation information set in the mimicry environment at the time t, and the propagation link of the topic community can act on other users.

Whereas the SFI-PE model is constructed by differential equations, the principal equations are as follows:

in the dynamics system constructed by the model, for the information propagation link depending on the social relationship, an active forwarding user has the average influence beta per unit time _F The capacity of N users is that at the current time t, the proportion of susceptible users in the total population is S (t)/N, so that the users in the forwarding state contact beta per unit time _F S (t) susceptible users who will have p _F The average forwarding probability of (d) yields a forwarding behavior. Thus, there is β _F S (t) F (t) the vulnerable users are affected by the active forwarding users. Wherein p is _F β _F S (t) F (t) the user generates the forwarding behavior to be converted into the forwarding state (F), (1-p) _F )β _F S (t) F (t) the user is not involved in information dissemination because the user is not interested in the information and changes to the immune state (I). Over time, there is alpha _F F (t) the user no longer has the ability to influence other users, and the immune state (I) is changed. Under the algorithm mechanism of the social media platform, the forwarding information of one forwarding user has the average exposure probability of gamma, so that gamma F (t) pieces of information are screened and presented in the mimicry environment every unit time. Thus, similarly, for an information propagation link that relies on a mimicry environment, active information in a mimicry environment will contact β per unit time _B S (t) susceptible users who will have p _B The average forwarding probability of (a) yields the forwarding behavior. Thus, there is p _B β _B S (t) B (t) A user generates a forwarding behavior to be converted into a forwarding state (F), (1-p) _B )β _B S (t) B (t) the user is not involved in information dissemination because the user is not interested in the information and changes to the immune state (I). Over time, α _B B (t) pieces of information no longer have the ability to influence other users and arriveThe end of the information lifecycle.

In the SFI-PE model, a social media information propagation system under the information ecological sight threshold is constructed. Next, the invention further considers the intervention effect of the third-party measures outside the system on the system to construct a third-party intervention information propagation dynamic model based on the social media mimicry environment, namely a macro-controlled stable-forwarding-immune pseudo-environment (M-SFI-PE) model.

The M-SFI-PE model is constructed on the basis of the SFI-PE model by considering the influence of each link of a third-party intervention system on information propagation, and the model architecture of the M-SFI-PE model is shown in figure 1 b. Third party intervention is not considered from this perspective, as the average immunization rate of the forwarding user and the average metabolic rate of the information in the mimicry environment are generally closely related to the usage habits of the individual user. In the M-SFI-PE model, the original assumption premise of the SFI-PE model is still kept.

The new parameters involved in the M-SFI-PE model and the corresponding explanation are shown in the following table 2, wherein beta _F 、β _B 、p _F 、p _B The parameters γ, etc. are defined as shown in table 1 above.

TABLE 2M-SFI-PE model parameter Table

It is worth noting that the effect of the third party intervention is related to the sign of the above parameters, i.e. the third party intervention can either suppress the propagation of the information backwards or promote it forwards. Considering that the M-SFI-PE model is established on the basis of the SFI-PE model, the main equation is as follows:

similar to the SFI-PE model, after considering the impact of adding third party intervention, it can be considered that: in the dynamic system provided by the invention, information propagation depending on social relations is realizedFor a link, an active forwarding user has an average impact per unit time (β) _F -β _MF ) The capability of N users, at the current time t, the proportion of susceptible users in the total population is S (t)/N, so that the users in the forwarding state will contact (beta) per unit time _F -β _MF ) S (t) susceptible users who will have (p) _F -p _MF ) The average forwarding probability of (a) yields the forwarding behavior.

Therefore, there is (β) _F -β _MF ) S (t) f (t) the vulnerable users are affected by the active forwarding user. Wherein (p) _F -p _MF )(β _F -β _MF ) S (t) F (t) the user generates the forwarding behavior to be converted into the forwarding state (F), (1-p) _F +p _MF )(β _F -β _MF ) S (t) F (t) the user is not involved in information dissemination because the user is not interested in the information and changes to the immune state (I). With the passage of time, there is alpha _F F (t) the user no longer has the ability to influence other users, and the immune state (I) is changed. Under the algorithm mechanism of the social media platform, the forwarding information of one forwarding user has (gamma-gamma) _M ) So there will be (gamma-gamma) per unit time _M ) F (t) pieces of information are screened and presented in the mimicry environment.

Thus similarly, for an information propagation link that relies on a mimicry environment, active information in a mimicry environment will be in contact (β) per unit time _B -β _MB ) S (t) susceptible users who will have (p) _B -p _MB ) The average forwarding probability of (a) yields the forwarding behavior. Thus, there is (p) _B -p _MB )(β _B - β _MB ) S (t) B (t) A user generates a forwarding behavior to be converted into a forwarding state (F), (1-p) _B +p _MB )(β _B - β _MB ) S (t) B (t) the users do not participate in information dissemination because the users are not interested in the information and are converted into the immune state (I). Over time, α _B The B (t) piece of information no longer has the ability to affect other users, and the end of the information life cycle is reached.

After the model is created, the model can be applied to information transmission analysis. The information propagation analysis method based on the social media mimicry environment modeling is used for carrying out information propagation analysis on the basis of the information propagation dynamic model based on the social media mimicry environment modeling.

FIG. 2 illustrates a flow diagram of a method for information dissemination analysis based on social media mimicry environment modeling, according to an embodiment of the invention.

As shown in fig. 2, the method for information dissemination analysis based on social media mimicry environment modeling provided by this embodiment includes the following steps:

s210: performing information transmission analysis by using an information transmission dynamic model based on the social media mimicry environment;

s220: and performing information transmission analysis on the basis of the information transmission dynamic model based on the social media mimicry environment in combination with a third-party intervention information transmission dynamic model based on the social media mimicry environment.

Wherein, step S210 further includes:

s211: searching and collecting original information data of an event to be analyzed on a preset social platform based on a preset search engine, wherein the information data comprises a user forwarding text and forwarding time; the event to be analyzed is a specific example of information, and the user forwarding text is a text forwarded by all forwarding individuals in the propagation group.

S212: preprocessing original information data to obtain a noiseless redundant user, a forwarding text and forwarding time under each piece of information;

s213: carrying out data fitting and parameter estimation by taking the noiseless redundant user forwarding text and the forwarding time as data driving; estimating parameters and initial susceptible population total of the information propagation dynamic model based on the social media mimicry environment by using a least square method;

s214: and determining information propagation index data of the event to be analyzed through the information propagation dynamic model based on the social media mimicry environment based on the parameters and the initial susceptible population total number.

In order to determine the information disseminationThe method constructs an information propagation index for expressing the development condition of information in the information propagation process. The information dissemination index includes an information dissemination reproducible number

Information propagation peak, information propagation final scale, and information propagation climax time.

Wherein the information can be propagated and reproduced

Used for judging whether the information is possible to spread in an outbreak; the information transmission peak value is used for measuring the highest point of public opinion outbreak; the final scale of information propagation is used for measuring the range to which the information propagation can be finally diffused; the information transmission climax time is used for measuring the speed of public opinion outbreak to the highest point. The method for acquiring the information propagation index data in the present invention will be described in further detail below.

Information dissemination reproducible number

In the model of an infectious disease,

is a substantially reproducible number and represents the average number of secondary infections of one patient during the average infection period. Similarly, in the SFI-PE model of the invention, the basic reproducible number is extended to represent the average value of the secondary forwarders caused by the effective information in each forwarding user and each mimicry environment under the condition of excluding the external intervention and all users being susceptible to influence

It is determined whether the information is likely to be bursty.

At the beginning of information distribution, if the forwarding amount per unit time is decreased, public opinion is not exploded. I.e. in the preceding formula(1) When S (t) is S ₀ When it is satisfied

Public opinion shows a trend of decay, namely:

since the forwarding group number F (t) and the effective propagation information number B (t) in the social media mimicry environment are non-negative numbers, the following conditions are met:

F(t)≥B(t)>0 (4)

therefore, the above formula (3) can be converted into

Thereby obtaining information propagation reproducible number

wherein ,

the method means that at the beginning of information release, the total number of forwarding groups shows a descending trend, and public opinion never explodes;

it means that at the beginning of information distribution, the total number of forwarding groups shows exponential growth, public opinion inevitably explodes, and

the larger the burst, the faster the burst rate.

Information propagation peak, information propagation final scale, and information propagation climax time:

in the information propagation process, the accumulated forwarding amount can be used for representing the propagation scale of the information. From the above differential equation (1), further derived, a differential equation of the accumulated forwarding amount shown in the following equation (7) can be derived:

in order to fit the real data to the model, the invention estimates the model parameters and the initial susceptible population total by the least squares method. Setting the parameter vector as theta ═ beta _F ,β _B ,p _F ,p _B ,α _F ,α _B ,γ,S ₀ ) By f _C (k, Θ) represents an analog value of the accumulated transfer amount at time k, and is represented by C _k And accumulating the true value of the forwarding amount in time. From this, a least squares LS error function can be obtained:

where LS is the sum of the squared residuals, k is 0,1,2, …, and T represents the sampling time. In the data fitting process, the parameters need to satisfy the following conditions: 1-p _F ≥0、1-p _B ≥0、1-γ≥0。

The accumulated forwarding amount of the event is a curve changing along with time, the general trend firstly rises rapidly and then rises steadily, and finally tends to be stable; the instantaneous forwarding of an event is a bell-shaped curve that rises first and then falls as time increases. In the present invention, the maximum value of the bell-shaped curve characterizing the instantaneous forwarding quantity is defined as the information propagation peak value (F) _max ) The method is used for measuring the highest point of public opinion outbreak, namely the information transmission strength; at the same time, the maximum value of the cumulative forwarding amount C (t) is defined as the final scale (C) of information propagation _s ) And is used for measuring the range to which the information transmission can be finally spread, namely the information transmission breadth.

In addition, in the present invention, the time to reach the information propagation peak is defined as the information propagation climax time

The method is used for measuring the speed of public opinion outbreak to the peak.

The present invention will be described in more detail below with reference to a study example of network information dissemination and process using the present invention.

According to the method, the latest information events are widely searched on the social platform, a popular theme is selected on the basis of a preset social platform (a Xinlang microblog platform in China is adopted in the embodiment), the information propagation dynamics mode of the social media is analyzed from a brand-new information ecological visual threshold, and the internal interaction between people and the internal interaction between environments in the system are researched.

In order to further study the influence of external factors on information propagation by each link of a third-party intervention system, the method screens out a positive typical event and a negative typical event under the theme of the heat, and collects real information data containing forwarding texts and accurate forwarding time under target information. Since the user mainly browses information during the physiological activity period and stops browsing information during the sleep time, in this embodiment, the collected raw information data needs to be preprocessed first, and the raw information data is filtered to avoid information stagnation caused by physiological requirements. After preprocessing, the noiseless instantaneous forwarding time point and forwarding text under each piece of information can be obtained, and therefore the accumulated forwarding amount is calculated.

In terms of obtaining the accumulated forwarding amount, the present embodiment obtains the accumulated forwarding amount by adding the number of users corresponding to the instant time point within a certain time range. Here, the start time is set to 0, and the sampling frequency is set to 1 hour. Next, in order to obtain the best fitting result, the present embodiment uses the true accumulated forwarding amount of the user as data driving, and estimates the SFI-PE model parameters and the initial susceptible population total by using the least square method described above.

Fig. 3 shows information propagation curves for a specific positive event (left) and a specific negative event (right) according to an embodiment of the invention. As shown in fig. 3, asterisks indicate the true accumulated forwarding amounts, and solid lines indicate the simulation values of the accumulated forwarding amounts calculated by the SFI-PE model. From the numerical simulation result, the fitting curve of the SFI-PE model is almost coincident with the real data point, so that the SFI-PE model is proved to be capable of fully representing the information transmission process, and the effectiveness of the model is verified. In this embodiment, the positive event parameter estimation value and the negative event parameter estimation value obtained by using the actual user forwarding amount as data driving are shown in tables 3 and 4 below, respectively.

TABLE 3 Positive event parameter results for SFI-PE model

TABLE 4 negative event parameter results for the SFI-PE model

As can be seen from the above two tables, the initial susceptible number S of positive events ₀ 8000, initial susceptible number of people S for a negative event ₀ 80000. For positive and negative events, the information forwarded by the forwarding user is screened by the platform algorithm mechanism to present the average exposure probability γ in the mimicry environment in a relatively small and approximately equivalent value, the positive event is γ ═ 0.1000, and the negative event is γ ═ 0.1040, which shows that only a very small number of information forwarded by the forwarding user can be screened and presented in the mimicry environment under the influence of the platform algorithm mechanism. Average contact rate beta related to network structure _F and β_B The method is stable on a small magnitude, and because the user can directly contact information on a personal homepage through social relations and needs to actively enter pages such as topic communities and the like through mimicry environment contact information, the method has the advantages that the beta value is small, the social relation is small, and the user can contact the information on the personal homepage, the user can enter the pages such as topic communities and the like _F Numerical ratio of (B) _B And is larger. Average forwarding probability p generated after vulnerable users contact information through social relations _F Less than the average forwarding probability p generated after contacting the information through social relations _B Showing that susceptible users are passing through the mimicry ringThe contact with the information is preferably followed by forwarding the information. Average immunization Rate alpha of the user _F Usually related to its own behavior law, and the average metabolic rate a of the information _B Generally, the average metabolic rate α of the information is found in relation to the degree of heat of the event and social concern _B Significantly less than the average immune rate a of the user _F It is shown that the exposure period during which the information is active in the mimicry environment is longer, and the time for affecting the information propagation is also longer. By using the model numerical fitting result, we can calculate specific numerical values of the event information propagation indexes, and the index numerical values of the positive events and the negative events are specifically shown in tables 5 and 6 below.

TABLE 5 information dissemination indicator results for specific positive events

TABLE 6 information dissemination indicator results for specific negative events

It can be seen that the information propagation of the positive and negative events can be repeated

At the beginning of information release of two types of events, the total number of forwarding groups is exponentially increased, so public opinion is inevitably broken out. Since two specific events studied by the present embodiment are social security issues surrounding information attention, the information propagation index results thereof are compared in the present embodiment. First, the information propagation peak F of the negative event _max Much larger than the information propagation peak F of the positive event _max The information propagation strength of the negative event in the information propagation process is larger than that of the positive event. Second, the time required to reach the information propagation peak for the negative event

Greater than the time required to reach the information propagation peak of the positive event

Indicating that the positive event first reaches the peak of the public opinion outbreak. Finally, the information propagation of the negative event is of final size C _s Information propagation final scale C much larger than the positive event _s The information spread of the negative event is wider.

In the SFI-PE model of the invention, the population S initially susceptible ₀ Is a variable to be estimated, which can determine the variation of the index, so this embodiment is considered as a parameter to be analyzed and experimented with other parameters preset by the SFI-PE model. In order to further explore the influence of parameter variation on the index, the embodiment adopts a correlation coefficient of deviation (PRCCs) method, which performs repeated experiments within the boundary range of the input parameters by using 1000 groups of samples, and finally gives the average sensitivity result of each parameter. The result of the method is between-1 and 1, if the result is close to 1 (the general trend of the scatter diagram is rightward), the strong positive influence of the input parameters on the information propagation index is shown, and if the result is close to-1 (the general trend of the scatter diagram is leftward), the strong negative influence of the input parameters on the information propagation index is shown. To explore the parameters (. beta.) _B 、β _F 、p _B 、p _F 、γ、α _B 、α _F 、S ₀ ) For information propagation peak F _max Information dissemination Final Scale C _s And information propagation reproducible number

In this embodiment, the PRCC result is visually represented by a histogram and a scatter diagram having a correspondence relationship, taking a positive event as an example.

FIG. 4 shows an index F under the influence of various parameters according to an embodiment of the invention _max 、C _s 、

PRCC junction ofSchematic representation of the fruit. Information propagation peak value F in the invention _max The strength of information propagation is characterized, and as can be seen from FIG. 4, the average contact rate beta at which a vulnerable user can contact information through social relations _F Average forwarding probability p generated after the susceptible user contacts the information through the mimicry environment _B Average forwarding probability p generated after susceptible users contact information through social relations _F Average immunization rate alpha of forwarding user _F And initial susceptible population total S ₀ Forwarding peak value F for index information _max The significance level of (a) is much less than 0.01, indicating that the correlation is very significant. Wherein p is _F and S₀ To F _max Having a strong positive correlation effect, beta _F To F _max Having a generally positive correlation effect, p _B To F _max Having a weak positive correlation effect, alpha _F To F is aligned with _max With a strong negative correlation effect.

Information dissemination Final Scale C _s Characterizing the breadth of information dissemination. As can be seen from FIG. 4, the average forwarding probability p generated after the vulnerable user contacts the information through the mimicry environment _B Average forwarding probability p generated after susceptible users contact information through social relations _F Average metabolic rate alpha at which information in a mimicry environment becomes inactive during information dissemination _B Average immunization rate alpha of forwarding user _F And initial susceptible population total S ₀ Propagation of the indicator information on the final scale C _s The significance level of (a) is much less than 0.01, indicating that the correlation is very significant. Wherein p is _F and S₀ To C _s Having a strong positive correlation effect, p _B To C _s Having a generally positive correlation effect, α _B and α_F To C _s With a weak negative correlation effect.

Information dissemination reproducible number

Characterize whether public opinion will break out. As can be seen from FIG. 4, the average contact rate β at which a vulnerable user can contact information through a mimicry environment _B Easy to useAverage contact rate beta at which an affected user can contact information through social relationships _F Average forwarding probability p generated after susceptible users contact information through mimicry environment _B Average forwarding probability p generated after susceptible users contact information through social relations _F Average immunization rate alpha of forwarding user _F And initial susceptible population total S ₀ Propagation of reproducible numbers to index information

The significance level of (a) is much less than 0.01, indicating that the correlation is very significant. Wherein p is _F and S₀ To R ₀ Having a strong positive correlation effect, beta _B 、β _F and p_B For is to

Having a generally positive correlation effect, α _F For is to

With a strong negative correlation effect.

The above results show that the parameter p is increased _F 、S ₀ 、β _F 、p _B While reducing alpha _F The method is favorable for improving the information transmission intensity, and is favorable for reducing the information transmission intensity; increasing the parameter p _F 、S ₀ and p_B While reducing alpha _B and α_F It is advantageous to enlarge the scale of information dissemination and, conversely, to reduce the scale of information dissemination. Increasing the parameter p from the perspective of whether public opinion is exploding _F 、S ₀ 、β _B 、β _F and p_B While reducing alpha _F Is beneficial to promoting public opinion outbreak, and conversely is beneficial to inhibiting public opinion outbreak.

Taking into account beta _F 、β _B 、p _F 、p _B Are important parameters influencing information propagation, and it is necessary to further study the specific influence of the parameters on each variable (instantaneous forwarding amount f (t), accumulated forwarding amount c (t)) for representing propagation trend and effective propagation information set b (t) in a mimicry environment.Fig. 5 shows a schematic diagram of information propagation indicator fluctuation caused by a single parameter change according to an embodiment of the invention.

As shown in FIG. 5, from the parameter fitting results of the SFI-PE model, the average contact rate β at which the susceptible user can contact the information through the social relationship is known _F ＝9.4000×10 ^-4 Average contact rate beta at which a vulnerable user can contact information through a mimicry environment _B ＝4.1900×10 ^-4 Average forwarding probability p generated after the vulnerable user contacts the information through social relation _F 0.4610, average forwarding probability p generated after a vulnerable user contacts information through a mimicry environment _B 0.8500, the information forwarded by the forwarding user is screened by a platform algorithm mechanism to show that the average exposure probability gamma in the mimicry environment is 0.1000, and the average immunization rate alpha at which the forwarding user becomes inactive in the information propagation process _F 1.5274, the average metabolic rate α at which information in a mimicry environment becomes inactive during information dissemination _B 0.1080, total number of initially susceptible people S ₀ 8000. By varying the parameter beta separately _F 、β _B 、 p _F 、p _B Meanwhile, other parameters are kept unchanged, and information propagation index change caused by single parameter change can be researched.

By observing (a) and (b) in fig. 5, it can be found that, for the curve of the cumulative forwarding amount c (t), the average contact rate β at which the vulnerable user can contact the information through social relations _F The reduction of (1) makes the curve increase speed and slow down in the early stage, but because the duration of the peak value of the effective propagation information set B (t) in the mimicry environment is prolonged, the curve slowly increases in the later stage, and therefore the final scale C of information propagation _s On the contrary, the average contact rate beta is slightly enlarged, and the user who is easily influenced by the counter-observation can contact the information through the mimicry environment _B The reduction of (A) has very weak influence on the early stage of the curve, and has very weak influence on the later stage of the curve and the final scale C of information transmission _s Has obvious inhibiting effect. For the curve of the instantaneous forwarding quantity F (t), the parameter β _F and β_B All make the information propagation peak value F _max Is decreased by _F The effect of (a) is relatively more pronounced. For the curves of the set of valid propagation information B (t) in the mimicry environment, the parameter β _F The reduction of (b) causes the effective propagation information peak in the mimicry environment to be decreased and the peak arrival time to be shifted, while beta _B The effect of (c) is very weak.

By observing (c) and (d) in fig. 5, it can be found that, for the curve of the cumulative forwarding amount c (t), the average forwarding probability p generated after the susceptible users contact the information through social relations _F The influence of the reduction on the curve is in the whole process, the curve is obviously accelerated and slowed down, and the final scale C of information transmission _s Obviously reducing the average forwarding probability p generated after the user which is easily influenced by counter-observation is exposed to the information through the mimicry environment _B The influence of the reduction on the curve is mainly concentrated in the later period, and the influence is relatively weak. For the curves of the instantaneous forwarding quantity F (t) and the set of valid propagation information B (t) in the mimicry environment, the parameter p _F Is reduced so that the information propagation peak F _max The peak value of the effective propagation information in the mimicry environment is obviously reduced, and p is _B The effect of (c) is very weak.

Through the analysis, the influence of the parameters on the information propagation indexes can be more intuitively understood, so that the following conclusion can be obtained:

first, the parameters (β) associated with the propagation link of the mimicry environment _B ，p _B ) For information propagation peak F _max The effect of (a) is not significant; second, the parameters (β) associated with the propagation link of the mimicry environment _B ，p _B ) The influence on the information propagation scale is mainly concentrated in the later period; third, a parameter (β) related to the social relationship of the propagation link _F ，p _F ) For information propagation peak F _max The influence of (a) is more significant; fourth, a parameter β related to the social relationship of the propagation link _F The influence on the information propagation scale is mainly concentrated in the early stage, and the final propagation scale C is expanded in the later stage _s And p is _F The impact on the information dissemination scale is global.

Therefore, the four parameters can be intervened to influence the information propagation trend, and the information propagation strength and the information propagation width are adjusted, so that the effective macroscopic regulation and control of the information propagation are realized. In addition, the average exposure probability gamma of the information forwarded by the forwarding user, which is screened and presented in the mimicry environment through the platform algorithm mechanism, reflects the function of the social media platform algorithm mechanism, and has an important significance in the research of the parameter gamma on the third-party intervention of information propagation.

And taking the two typical events as basic scenes, taking the parameter values estimated by the SFI-PE model as a comparison group, and carrying out moderate assumption on the newly added parameters of the M-SFI-PE model according to the parameter sensitivity analysis result of the information propagation index. By applying a control variable method, 5 new scenes are designed to form an experimental group in the embodiment, so as to simulate the influence of different third-party intervention means on the crowd state and behavior in the social media information ecosystem and investigate the direction which should be selected by the third-party intervention information propagation.

First, the intervention study on the third party to facilitate the dissemination of positive event information is as follows:

according to the parameter fitting result of the SFI-PE model, the average contact rate beta of the susceptible users which can contact the information through the social relation can be known _F ＝9.4000×10 ^-4 Average contact rate beta at which a vulnerable user can contact information through a mimicry environment _B ＝4.1900×10 ^-4 Average forwarding probability p generated after the vulnerable user contacts the information through social relation _F 0.4610, average forwarding probability p generated after the vulnerable user is exposed to the information through the mimicry environment _B 0.8500, the average exposure probability gamma of the forwarded information of the forwarding user is screened by a platform algorithm mechanism to be presented in the mimicry environment is 0.1000, and the average immunity rate alpha of the forwarding user which becomes inactive in the information transmission process _F 1.5274, average metabolic rate α at which information in a mimicry environment becomes inactive during information dissemination _B 0.1080, total number of initially susceptible people S ₀ 8000. On the basis, the specific setting of the newly added parameters of the M-SFI-PE model for promoting the propagation of the positive event information by the third party is shown in the following table 7.

TABLE 7 third M-SFI-PE model Add-on parameter setup to facilitate propagation of Positive event information

The parameter values in table 7 are all set to 1/10 for the intervened parameter, and the minus sign before the value reflects the promotion effect of the third party intervention on information propagation. For example, in setting the newly added parameters of the M-SFI-PE model, β _MF ＝-9.4000×10 ^-5 Value of beta _F ＝9.4000×10 ^-4 1/10, minus sign "-" indicates a promoting effect. And performing an experiment for promoting information propagation by considering third-party intervention on the event, and performing simulation analysis on the M-SFI-PE model by using Matlab, thereby obtaining the number S (t) of the susceptible state, the number F (t) of the forwarding state, the number I (t) of the immune state and the change condition of the effective propagation information set B (t) in the mimicry environment in each scene.

Fig. 6 is a schematic diagram illustrating the crowd evolution of a third party promoting the propagation of positive event information according to an embodiment of the present invention. As shown in fig. 6, in the simulation result for promoting the propagation of the frontal event information, the control group and the experimental group (a) β are respectively set in the order of the line marks in fig. 6 from top to bottom _MF ＝-9.4000×10 ^-5 Situation of (a), experimental group (b) beta _MB ＝-4.1900×10 ^-5 Situation of (1), Experimental group (c) p _MF Scenario, experimental group (d) p ═ 0.0461 _MB Scenario-0.0850, experimental group (e) γ _M Scene-0.0100. As can be seen from fig. 6 (a) and (b), in the conversion of the susceptible state (S) to the forwarding state (F), only the experimental group (c) helps to accelerate the decrease in the number of susceptible state people S (t). Experimental groups (a) and (c) may increase the peak in the population f (t) of the forwarding state, so that the peak point of information propagation comes faster. While other intervention methods are relatively ineffective. As can be seen from the observation of (b) and (c) of FIG. 6, in the transition from the forwarding state (F) to the immune state (I), the forwarding user does not become inactive on average during the information dissemination processRate of immunization alpha _F By direct intervention, at p _MF Under the influence of (2), the population number I (t) in the immune state is synchronously increased, but the final immune population scale is not obviously enlarged, so that the proportion of effective information dissemination population is improved. As can be seen from the observation of the graph (d), the five intervention modes all contribute to increasing the effective information amount in the mimicry environment, wherein the parameter p is adjusted _MF and γ_M The mode has the best intervention effect, the peak value of the effective information of the achieved mimicry environment is the highest, and the descending rate is obviously slowed down.

And then, specific numerical values of the information propagation indexes after intervention can be calculated by using the new model parameter values, so that the five intervention modes are evaluated. The information propagation index values after the positive event intervention are specifically shown in table 8 below.

TABLE 8 information dissemination indicator results for specific positive events after intervention

By comparing the results of the indicators in table 8, the following conclusions can be drawn:

1. from the information propagation peak F if it is desired to increase the intensity of the information propagation _max Viewed in index by adjusting the parameter p _MF The mode (2) has the optimal intervention effect, and the order is p on the intervention mode _MF >β _MF > p _MB >β _MB ≈γ _M (ii) a 2. From the final size C of the information dissemination, if it is desired to expand the breadth of the information dissemination _s Viewed in index by adjusting the parameter p _MF The mode (2) has the optimal intervention effect, and the order is p on the intervention mode _MF >p _MB >β _MB ≈γ _M By adjusting the parameter beta _MF Will instead reduce the information propagation final size C _s (ii) a 3. If the time of reaching the highest point of the public sentiment is expected to be faster, the information transmission climax time t _Fmax To see, the order is selected as p on the intervention mode _MF ≈β _MF >β _MB ≈p _MB ≈γ _M 。

In summary, the average forward probability intervention rate p generated after the susceptible users contact the information through the social relationship is regulated _MF The method has the best intervention effect in promoting the information dissemination of positive events, and the average forwarding probability p generated after the susceptible users contact the information through social relations _F Are key parameters that facilitate the propagation of positive event information. Therefore, for the information dissemination of the positive events, the effective direction of the third party intervening the information dissemination is to improve the forwarding willingness of the susceptible users to contact the information through social relations.

Secondly, intervention studies on third parties to suppress the propagation of negative event information are as follows:

the average contact rate beta of the susceptible users which can contact the information through social relations can be known from the parameter fitting result of the SFI-PE model _F ＝1.3400×10 ^-4 Average contact rate beta at which a vulnerable user can contact information through a mimicry environment _B ＝1.200×10 ^-4 Average forwarding probability p generated after susceptible users contact information through social relations _F 0.2200, average forwarding probability p generated after the vulnerable user is exposed to the information through the mimicry environment _B 0.7290, the information forwarded by the forwarding user is screened by a platform algorithm mechanism to present an average exposure probability gamma of 0.0104 in the mimicry environment, and the average immunization rate alpha of the forwarding user becoming inactive in the information dissemination process _F 1.3384, average metabolic rate α at which information in a mimicry environment becomes inactive during information dissemination _B 0.1740, total number of initially susceptible people S ₀ 80000. On the basis, the specific setting of the M-SFI-PE model newly added parameters for inhibiting the propagation of the negative event information by the third party is shown in the following table 9.

TABLE 9 newly added parameter settings of M-SFI-PE model for third party to suppress propagation of negative event information

The values of the parameters in table 9 are all set to 1/10 for the intervened parameter, and the plus sign before the value reflects the inhibition of the third party intervention on the information propagation. For example, in setting the newly added parameters of the M-SFI-PE model, β _MF ＝1.3400×10 ^-5 Value of beta _F ＝1.3400×10 ^-4 1/10, positive sign "+" indicates inhibition. And carrying out an experiment for inhibiting information propagation by considering third-party intervention on the event, and carrying out simulation analysis on the M-SFI-PE model by utilizing Matlab, thereby obtaining the change conditions of the population quantity S (t) in the susceptible state, the population quantity F (t) in the forwarding state, the population quantity I (t) in the immune state and the effective propagation information set B (t) in the mimicry environment in each scene.

Fig. 7 is a schematic diagram illustrating the crowd evolution of the third party suppressing the propagation of the negative event information according to the embodiment of the present invention. As shown in fig. 7, in the simulation result of the third party inhibiting the propagation of the negative event information, the results are respectively the control group and the experimental group (a) β according to the sequence of the line shape marks in fig. 7 from top to bottom _MF ＝1.3400×10 ^-5 Situation of (c), Experimental group (b). beta _MB ＝1.200×10 ^-5 Situation of (1), Experimental group (c) p _MF Situation, experimental group (d) p of 0.0220 _MB Scenario, experimental group (e) γ of 0.0729 _M Scene 0.0104. As can be seen from observing (a) and (b) in fig. 7, in the process of converting the susceptible state (S) to the forwarding state (F), the experimental groups (a) and (c) help to slow down the decrease of the population number S (t) in the susceptible state, and reduce the peak of the population number F (t) in the forwarding state, so that the peak point of information propagation is delayed, and the significance p of the intervention effect is _MF >β _MF . As can be seen from the observation of (b) and (c) in fig. 7, in the process of converting the forwarding state (F) to the immune state (I), the average immune rate α at which the forwarding user does not become inactive during the information dissemination process is not used _F By direct intervention, at p _MF and β_MF The population number i (t) in the immune status is reduced, but the final immune population size is not significantly reduced, and the proportion of the effective information dissemination population is reduced. As can be seen from (d) in FIG. 7, five kinds of the compoundsThe intervention mode is helpful to reduce the effective information quantity in the mimicry environment, wherein the parameter gamma is adjusted _M The mode (2) has the best intervention effect, the peak value of the obtained effective information of the mimicry environment is obviously reduced, and then the parameter p is adjusted _MF Again, adjusting the parameter p _MB and β_MB By adjusting the parameter beta _MF The intervention effect is least obvious.

And then, calculating a specific numerical value of the information propagation index after intervention by using the new model parameter value, thereby evaluating the five intervention modes. The information propagation index values after the adverse event intervention are specifically shown in table 10 below.

TABLE 10 information dissemination indicator results for specific negative events after intervention

By comparing the results of the indices in table 10, the following conclusions can be drawn:

1. from the information propagation peak F if it is desired to reduce the intensity of the information propagation _max In terms of index, by adjusting the parameter p _MF The mode (2) has the optimal intervention effect, and the order is p on the intervention mode _MF >β _MF > p _MB >β _MB ≈γ _M (ii) a 2. From the information dissemination final scale C if it is desired to reduce the breadth of information dissemination _s In terms of index, by adjusting the parameter p _MF The mode (2) has the optimal intervention effect, and the order is p on the intervention mode _MF >p _MB >β _MB ≈γ _M By adjusting the parameter beta _MF Instead, the information propagation final size C is increased _s (ii) a 3. If the time reaching the highest point of public sentiment is hoped to be delayed, the climax time is propagated from the information

To see, the order p is selected on the intervention mode _MF ≈β _MF >β _MB ≈p _MB ≈γ _M 。

The conclusions reached by the above embodiments are mutually corroborated with third party interventions to facilitate the information dissemination experimental results of the proof events in terms of third party interventions to suppress information dissemination of negative events. That is, in general, the average forward probability intervention rate p generated after the vulnerable users contact the information through the social relationship is regulated by the third party _MF The method has the best intervention effect in the aspect of inhibiting the information propagation of negative events, and the average forwarding probability p generated after the susceptible users contact the information through social relations _F Is a key parameter for suppressing the propagation of negative event information. Therefore, for the information dissemination of the negative event, the effective direction for intervening the information dissemination by the third party is to reduce the forwarding willingness of the vulnerable users to contact the information through the social relationship.

Through the above-described embodiments, the validity of two models (SFI-PE model and M-SFI-PE model) can be verified. On one hand, the SFI-PE model under the visual threshold of the information ecosystem is constructed by adopting the research thought of the system theory. By establishing a novel network society construction mechanism, two kinds of propagation links of social media platform information propagation are deeply analyzed, namely social relations constructed by platform mechanisms such as 'vermicelli', 'concern' and the like and mimicry environments constructed by platform mechanisms such as 'hot', 'keyword search' and the like, the interaction and dynamic influence between people and between people in a social media information ecosystem and between people and the environment are comprehensively researched, and the general rule of the network information propagation ecosystem under the current new media environment can be summarized by combining parameter sensitivity analysis on information propagation indexes. On the other hand, the influence of each link of a third-party intervention system on information propagation is further considered on the basis of the SFI-PE model, and the M-SFI-PE model under the action of third-party intervention considering the external factors of the system is constructed. The embodiment develops a control experiment aiming at the promotion of the propagation of the positive event and the inhibition of the propagation of the negative event by focusing on the hot spot problem and taking two typical events as examples, thereby obtaining an effective method and conclusion of information propagation intervention of third party intervention. In general, whether to facilitate or inhibit the propagation of positive events or negative events, are susceptible toThe average forwarding probability p generated after the user contacts the information through the social relationship _F Is the key cut-in point and the optimal direction of the third-party intervention.

The experimental conclusion is helpful for constructing an efficient information propagation intervention method. If the information propagation of positive events is to be promoted, the comprehensive optimal scheme is that the average forwarding probability p generated after the information is contacted by susceptible users through social relations through the intervention of a third party _F For example, more emotional value is added in the formulation of media issues, realistic scenes for implementing epidemic prevention and control and social security measures are shown by a method of combining pictures and texts and describing concierge, and people mostly use 'us' as a main language to enhance emotional resonance and make people have more participation and flu by calling for calls and other ways. If the information transmission of the negative events is to be inhibited, the comprehensive optimal scheme is that the average forwarding probability p generated after the information is contacted by the vulnerable users through social relations due to the intervention of a third party _F For example, for negative false information, a popup window can be set for the false information so that the contacted person can know the distortiveness of the information; for a real negative event, a third party can indicate that the third party knows about the relevant situation by adding a relevant topic, and because the spontaneous forwarding of the information is usually for getting attention in this situation, a jump link can be set on the page of the original negative information, so that a person who contacts the information later can directly click and jump to an information page responded by the third party. Therefore, the information propagation intervention method constructed by the invention can provide policy support for information governance so as to assist in maintaining social stability and network security.

Corresponding to the information propagation analysis method based on the social media mimicry environment modeling, the invention also provides an information propagation analysis system based on the social media mimicry environment modeling, which comprises a modeling unit and an information propagation analysis unit.

The modeling unit is used for creating an information propagation dynamic model based on the social media mimicry environment and a third-party intervention information propagation dynamic model based on the social media mimicry environment; the information propagation dynamic model based on the social media mimicry environment is an information propagation dynamic model based on the social media mimicry environment, which is established by utilizing the information propagation dynamic model modeling method based on the social media mimicry environment; the information propagation analysis unit comprises a primary analysis unit and an intervention analysis unit, wherein the primary analysis unit is used for carrying out information propagation analysis by utilizing the information propagation dynamic model based on the social media mimicry environment; the intervention analysis unit is used for carrying out information transmission analysis on the basis of the information transmission dynamic model based on the social media mimicry environment by combining a third-party intervention information transmission dynamic model based on the social media mimicry environment;

further, the primary analysis unit includes:

and the index data determining unit is used for determining the information propagation index data of the event to be analyzed through the information propagation dynamic model based on the social media mimicry environment based on the parameters and the initial susceptible population total number.

The information propagation analysis system based on the social media mimicry environment modeling corresponds to the information propagation analysis method based on the social media mimicry environment modeling, and the specific implementation manner of the information propagation analysis system based on the social media mimicry environment modeling can refer to the description of the information propagation analysis method based on the social media mimicry environment modeling, and is not described one by one here.

The mixed information propagation dynamic model based on individual emotional interaction and the method for information propagation analysis by applying the model according to the invention are described above by way of example with reference to the accompanying drawings. However, it should be understood by those skilled in the art that various modifications can be made to the above-mentioned mixed information dissemination dynamic model based on individual emotional interaction and the method for information dissemination analysis by using the model without departing from the scope of the present invention. Therefore, the scope of the present invention should be determined by the contents of the appended claims.

Claims

1. A modeling method of information transmission dynamic model based on social media mimicry environment is characterized in that, in the information transmission dynamic model,

in the development process of a horological event, information propagation generated by depending on social relations and information propagation generated by depending on a mimicry environment are two parallel propagation links which are not interfered with each other and have different information metabolism modes; information is spread in a closed and stable environment, the total number N of people in the environment is unchanged, N is divided into three groups of a susceptible state S, a forwarding state F and an immune state I, and the condition that each individual in the group is unique at any time is assumed; the meaning of each state representation in the information propagation dynamic model is as follows:

forwarding state F: the individual in the state generates the forwarding behavior and has the capability of infecting the individual in the susceptible state to forward the information;

immune status I: the population in this state consists of two parts:

the individual in the forwarding state exceeds the active exposure period and no longer has the ability to affect others, thereby converting to the immune state; and the number of the first and second groups,

after contacting information, an individual in a susceptible state is directly converted into an immune state because the information is not subjectively interested;

the differential equation of the information propagation dynamics model is as follows:

wherein ,β_F Average contact rate at which information can be contacted by social relationships for vulnerable users; beta is a _B Average contact rate, p, at which information can be contacted by a vulnerable user through a mimicry environment _F Average forwarding probability, p, generated after the vulnerable users contact the information through social relations _B The average forwarding probability, alpha, generated after the susceptible user is exposed to the information through the mimicry environment _F To forward the average immunization rate at which the user becomes inactive during information dissemination, α _B Gamma is the average metabolic rate at which the information in the mimicry environment becomes inactive in the information propagation process, and gamma is the average exposure probability at which the information forwarded by the forwarding user is screened through a platform algorithm mechanism and presented in the mimicry environment.

2. The modeling method of the information propagation dynamical model based on the social media mimicry environment according to claim 1, wherein the influence of a third-party intervention system on information propagation is added on the basis of the information propagation dynamical model, and a third-party intervention information propagation dynamical model based on the social media mimicry environment is constructed, wherein the assumption premise of the information propagation dynamical model based on the social media mimicry environment is maintained, and the differential equation of the third-party intervention information propagation dynamical model based on the social media mimicry environment is as follows:

wherein ,β_MF Is a third party to beta _F Interference coefficient of beta _MB Is a third party to beta _B Interference coefficient of p _MF Is a third party pair p _F Interference coefficient of p _MB Is a third party pair p _B Interference coefficient of (a), gamma _M Is the intervention coefficient of the third party to gamma.

3. An information propagation analysis method based on modeling of a social media mimicry environment is used for carrying out information propagation analysis by utilizing an information propagation dynamic model based on the social media mimicry environment, wherein the information propagation dynamic model based on the social media mimicry environment is an information propagation dynamic model based on the social media mimicry environment, which is established by utilizing the information propagation dynamic model modeling method based on the social media mimicry environment in claim 1 or 2; the method comprises the following steps:

performing information transmission analysis by using the information transmission dynamic model based on the social media mimicry environment; and the number of the first and second groups,

performing information transmission analysis on the basis of the information transmission dynamic model based on the social media mimicry environment by combining with a third-party intervention information transmission dynamic model based on the social media mimicry environment;

carrying out data fitting and parameter estimation by taking the noiseless redundant user forwarding texts and the forwarding time as data driving; estimating parameters and initial susceptible population total of the information propagation dynamic model based on the social media mimicry environment by using a least square method;

4. The information dissemination analysis method based on social media mimicry environment modeling of claim 3 wherein the information dissemination index data comprises an information dissemination regenerable number

Information propagation peak value, information propagation final scale and information propagation climax time; wherein,

said information dissemination reproducible number

For judging whether the information is possible to spread in an explosion;

the information transmission peak value is used for measuring the highest point of public opinion outbreak;

the information propagation final scale is used for measuring the range to which the information propagation can be finally diffused;

the information transmission climax time is used for measuring the speed of public opinion outbreak to the highest point.

5. The information dissemination analysis method based on social media mimicry environment modeling according to claim 4 wherein the information dissemination reproducible number

The method comprises the following steps of:

at the beginning of information distribution, if the forwarding amount per unit time is decreased, the public opinion does not burst, and in the differential equation of the information propagation dynamic model, when S (t) is S ═ S ₀ When it is satisfied

The public opinion shows a trend of decay, and at this time,

because the forwarding group number F (t) and the effective propagation information number B (t) in the mimicry environment are non-negative numbers, the following conditions are met: f (t) ≧ B (t) >0, thereby obtaining:

obtaining the information propagation reproducible number

wherein ,

the total number of the forwarding groups at the beginning of information release shows a descending trend, and public opinion can never be outbreak;

it means that at the beginning of information distribution, the total number of forwarding groups grows exponentially, public opinion inevitably explodes, and

the larger the outbreak, the faster the outbreak.

6. The information propagation analysis method based on modeling of social media mimicry environment of claim 5, wherein in the process of estimating the parameters of the information propagation dynamic model based on the social media mimicry environment and the initial susceptible population by using the least square method,

setting the parameter vector to theta ═ beta _F ,β _B ,p _F ,p _B ,α _F ,α _B ,γ,S ₀ ) By f _C (k, Θ) represents an analog value of the accumulated transfer amount at time k, and is represented by C _k And accumulating the true value of the forwarding quantity in time, thereby obtaining a least square LS error function:

wherein LS is the sum of squares of residuals, k is 0,1,2, …, and T represents the sampling time;

in the process of data fitting, the parameters of the information propagation dynamic model based on the social media mimicry environment need to satisfy the following conditions: 1-p _F ≥0、1-p _B ≥0、1-γ≥0。

7. The information dissemination analysis method based on social media mimicry environment modeling according to claim 6, wherein the physical quantities used to characterize the scale of information dissemination include an accumulated forwarding quantity and an instantaneous forwarding quantity, wherein the differential equation of the accumulated forwarding quantity is:

the accumulated forwarding amount is a curve changing along with time, the general trend firstly rises rapidly and then rises steadily, and finally tends to be stable; the instantaneous forwarding amount is a bell-shaped curve which rises firstly and then falls along with the increase of time; defining the maximum value of a bell-shaped curve characterizing the instantaneous forwarding quantity as the information propagation peak value F _max Defining the maximum value of the accumulated forwarding amount C (t) as the final information propagation scale C _s Defining the time of arrival at the information propagation peak as the information propagation climaxTime

8. An information dissemination analysis system based on social media mimicry environment modeling, comprising:

the modeling unit is used for creating an information propagation dynamic model based on the social media mimicry environment and a third-party intervention information propagation dynamic model based on the social media mimicry environment; the information transmission dynamic model based on the social media mimicry environment is an information transmission dynamic model based on the social media mimicry environment, which is established by the information transmission dynamic model modeling method based on the social media mimicry environment in claim 1 or 2;

the information propagation analysis unit comprises a primary analysis unit and an intervention analysis unit, wherein the primary analysis unit is used for carrying out information propagation analysis by utilizing the information propagation dynamic model based on the social media mimicry environment; the intervention analysis unit is used for carrying out information transmission analysis on the basis of the information transmission dynamic model based on the social media mimicry environment in combination with a third-party intervention information transmission dynamic model based on the social media mimicry environment;

wherein the primary analysis unit comprises:

9. The information dissemination analysis system based on social media mimicry environment modeling of claim 8 wherein the information dissemination index data comprises an information dissemination renewability number

said information dissemination reproducible number

Used for judging whether the information is possible to spread in an outbreak;

the information transmission climax time is used for measuring the speed of the public opinion outbreak to the peak.

10. The information dissemination analysis system based on social media mimicry environment modeling according to claim 8 wherein the physical quantities characterizing the size of information dissemination include an accumulated forwarding quantity and an instantaneous forwarding quantity, wherein the differential equation of the accumulated forwarding quantity is:

the accumulated forwarding amount is a curve changing along with time, the general trend rises rapidly at first and then steadily, and finally tends to be stable; the instantaneous forwarding amount isA bell-shaped curve rising first and falling second in time; defining the maximum value of a bell-shaped curve representing the instantaneous forwarding quantity as an information propagation peak value F _max Defining the maximum value of the accumulated forwarding amount C (t) as the final information propagation scale C _s Defining the time of reaching the information propagation peak as the information propagation climax time