An information propagation prediction and tracing method and system based on an iterative function system

Abstract

The application discloses an information propagation prediction and tracing method and system based on an iterative function system. The method comprises the following steps: firstly, obtaining information forwarding data of an event in a network platform, and constructing a basic network by using user data; secondly, performing affine transformation on the basic network by using an iterative function system, combining a SI infectious disease model, and constructing an information propagation fractal network according to an information propagation model; furthermore, analyzing the propagation mechanism of information according to the propagation model, and realizing prediction of the information propagation trend; finally, according to newly generated nodes and node edges of currently received information, the publisher of the information is reversely calculated by using an IFS code, and the information is traced. The application can predict and trace the information propagation, effectively control the propagation of public opinion information, and thus maintain the social stability. Meanwhile, the system has low memory cost, high operation efficiency, independence and portability.

Description

TECHNICAL FIELD

[0001] The present application relates to information propagation trend prediction and control technology, in particular to an information propagation prediction and tracing method and system based on an iterative function system. BACKGROUND

[0002] In the new media era, various online social platforms emerge in an endless stream, among which microblog attracts high participation of a large number of people. However, the current microblog has the following problems:

[0003] Firstly, the microblog provides a broad social platform for the public, but there is almost no restriction on the subject of information publishing, sharing and propagation. Due to the insufficient screening of information by the microblog, the information is mixed with true and false, leading to the widespread propagation of part of false and public opinion information, thereby affecting the social order and personal safety.

[0004] Secondly, the information distribution on the microblog is relatively scattered, and no clear and efficient aggregation mode is formed, so the required information cannot be accurately located and searched at the first time. Therefore, how to explore potential users and improve their attention and adhesion, and under the premise of aggregating scattered and potential information, has become a key problem to be solved by the microblog platform.

[0005] The complex microblog propagation path can be regarded as a generalized fractal structure. The fractal theory is used to study the characteristics of information propagation on the microblog, understand the user structure and user demand, and combine the technical analysis to visualize the complex information propagation path as a simple graph for description, so that the network media and network propagation can be understood from a new angle.

[0006] Iterative function system (IFS) is an important branch of fractal theory. It takes affine transformation as the framework, and according to the self-similar structure of the whole and the part of the geometric object, the overall shape is iterated according to different affine transformations with a certain probability until a satisfactory fractal graph is obtained.

[0007] At present, the iterative function system has been widely used in many fields, but its application in communication science still needs further research and exploration. SUMMARY

[0008] One object of the present application is to provide an information propagation prediction and tracing system based on an iterative function system. The second object of the present application is to provide an information propagation prediction and tracing method based on an iterative function system.

[0009] To achieve the above object, the information propagation prediction and tracing system based on the iterative function system comprises a data acquisition module, an initial network acquisition module, a propagation model module, an information aggregation module, an information propagation prediction module and an information tracing module.

[0010] The data acquisition module is configured to acquire first-level forwarding data of a certain event in a network platform, and the first-level forwarding data includes information data, forwarding user data, and opinion and comment data accompanying the forwarding of a user;

[0011] The initial network acquisition module is configured to perform denoising processing on the data acquired by the data acquisition module and construct a basic network;

[0012] The propagation model module includes an iterative function system module and an SI infectious disease model module;

[0013] The propagation model module is configured to generate an information propagation fractal network with fractal characteristics from the basic network constructed by the initial network acquisition module,

[0014] The iterative function system module is configured to generate IFS codes from the basic network constructed by the initial network acquisition module, and the IFS codes are composed of different forwarding layer network IFS codes,

[0015] The SI infectious disease model module is configured to generate edge rules of nodes in the network from the basic network constructed by the initial network acquisition module;

[0016] The information aggregation module is configured to aggregate scattered and potential information in the network, i.e., aggregate IFS codes of different forwarding layer networks;

[0017] The information propagation prediction module is configured to predict an information propagation path through the edge rules of nodes in the network;

[0018] The information tracing module is configured to trace the originator of information through the edge rules of nodes in the network, thereby realizing information tracing.

[0019] The information propagation prediction and tracing method based on the iterative function system includes the following steps:

[0020] Step 1: The first-level forwarding data of a certain event in a network platform is acquired through a crawler technology in the data acquisition module, and the acquired first-level forwarding data is used to establish a basic network through the initial network acquisition module, and the basic network is used as an affine change whole network;

[0021] Step 2: The basic network established in step 1 is projected onto a coordinate axis, wherein the coordinate position of a root user node is (0, 0), i.e., a central node, and an information propagation fractal network with fractal characteristics is generated by using the iterative function system module in the propagation model module;

[0022] Step 3: The information propagation prediction module uses the propagation dynamics equation of the SI infectious disease model module to depict the propagation process, thereby realizing prediction of information propagation;

[0023] Step 4: According to the coordinates of the final receiver of the public opinion information, the initial publisher of the public opinion information is reversely inferred by using the IFS iteration principle of the iteration function system module;

[0024] Step 5: Based on the opinion and comment data attached when the user forwards the message obtained in step 1, the iteration function system module (4) is used to aggregate the dispersed and potential information in the network, and finally an information fractal graph with complete information is generated; the public attitude and opinion are analyzed from the opinion and comment data attached when the user forwards the message, so that an event in the network platform is analyzed.

[0025] When the first-level forwarding data obtained in the step 1 is established by the initial network acquisition module, the initial publisher of the public opinion information is taken as the center node, the new node is the forwarding user data, and the node connection is the forwarding relationship.

[0026] The specific implementation steps of the step 2 are as follows:

[0027] Step 2.1: The root user node n0(0, 0) is taken as the initial publisher of the public opinion information to generate the first-level forwarding network; the IFS code affine variation set of the first layer is {R1 2 : ω1, ω1,.... ω m}, the probability set corresponding to the affine variation set is {p 11, p 12 ,.....p 1m}; from n0, the corresponding probability in the probability set {p 11 ,p 12 ,.....p 1m} is selected to generate the center node c_node1={n 11 ,n 12 ,....,n 1m ,} of the second layer network; the node n0 propagates information to the nodes in c_node1 with a probability δ;

[0028] Step 2.2: The nodes in c_node1 are taken as the initial publishers of the public opinion information in the next round of propagation, and each node in c_node1 is subjected to affine variation according to the IFS code affine variation set {R2 2 : ω1, ω1,.... ω m} of the second layer, and the corresponding probability {p 21 ,p 22 ,.....p 2m} is selected to generate the center node set c_node2={c_node 2i |i=1,...,m} of the subgraph, wherein c_node1 propagates information to c_node2i The nodes in c_node2 spread information to the nodes in c_node3, and if the spreading is successful, the nodes are connected.

[0029] Step 2.3: c_node2 is divided into m subgraphs, and the center nodes of the subgraphs are defined as c_node3={c_nodei|i=1,...,m}. 2i The nodes in c_node2 are defined as newborn nodes of a new round of infection, and each node in c_node2 is subjected to affine transformation set {R3: ω1, ω1,.... ω 2i} according to the IFS code of the third layer, and the corresponding probability {p 2 , p m ,.... p 31} is selected for radiation transformation to generate the center node set c_node3={c_nodei|i=1,...,m} of the subgraph, respectively. 32 3m 3i The nodes in c_node2 spread information to the nodes in c_node 3i , and if the transmission is successful, the nodes are connected.

[0030] The specific implementation steps of step 2.1 are as follows:

[0031] The fractal graph is drawn by the iterative function system, and a network is randomly determined as the whole of the geometric object through IFS code; the IFS code is calculated according to the determined network, different affine transformations are generated according to different IFS codes, thereby forming the set of compression affine transformations {R 2 : ω1, ω1,.... ω n}, and the probability set {p1, p2,.... p n} corresponding to the affine transformation set is obtained according to the area ratio of each subgraph, wherein p i >0 is the probability corresponding to the transformation, and satisfies

[0032] The affine transformation of the two-dimensional Euclidean space is defined as ω: R 2 → R 2 , (X, Y) is a point in the two-dimensional Euclidean space, and its affine transformation image is (X', Y'), which is written in matrix form as:

[0033]

[0034] The node is brought into the above expression:

[0035]

[0036] where ω represents the six parameters a, b, c, d, e, and f, and when a graph is divided into N local subgraphs, the set of compression mappings from the whole to the local is ω​​​n and its corresponding probability p n IFS code, denoted as {X:(ω n ,p n , n = 1, 2, 3... N}.

[0037] The propagation model module is specifically implemented as:

[0038] The nodes in a network are divided into two states, namely the infected state and the susceptible state, the infected person infects the susceptible person at a transmission rate of beta, and S(t) and I(t) represent the proportion of susceptible persons and infected persons at time t, and S(t) + I(t) = 1; Once the susceptible person is infected, he cannot be cured, and finally, with the passage of time t, everyone is infected; The propagation dynamics equation of the propagation model is:

[0039]

[0040] The propagation model is defined as follows:

[0041] The probabilities of IFS codes of different layers are defined as p i(i＝1,2,3) , wherein

[0042]

[0043] Where alpha is the initial infection rate of information; N is the total number of nodes in the network, p1 is the probability corresponding to the affine transformation of the first layer, p2 is the probability corresponding to the affine transformation of the second layer, and p3 is the probability corresponding to the affine transformation of the third layer.

[0044] The specific implementation of the step 3 is:

[0045] The differential equation of the SI infectious disease model module is a formula of Markov property, as shown in the following formula:

[0046]

[0047] The formula describes a random process of state transition to another state, which has the property of "no memory", that is, the probability distribution of the next state can only be determined by the current state, and it has nothing to do with the events before it in the time series. In the SI infectious disease model module, it is represented as the infection of the day is only related to the number of disease infections of the previous day, according to the change law of the propagation dynamics equation, the situation of the network and the information diffusion are judged, so as to predict the trend of information propagation.

[0048] The specific implementation of the step 4 is:

[0049] According to the current information receiving new node coordinates, according to the IFS code, the initial publisher of the public opinion information is obtained, and the information propagation is traced back.

[0050] The specific implementation of the step 5 is:

[0051] The first-level forwarding data and the IFS code generated by the iterative function system module in steps 2.1, 2.2 and 2.3 are iteratively generated to generate an information fractal network with complete information.

[0052] The network platform can also be a social media; the network platform is a microblog.

[0053] In step 2.1, the initial publisher of the public opinion information is taken as an initial "infector", and a propagation model is used to "infect" the new node generated based on the iterative function system according to the propagation probability, and if "infection" occurs, the node edge is established.

[0054] In step 3, based on the Markov property of the propagation dynamics equation, the iterative function system module is used to iteratively analyze the diffusion of information.

[0055] The IFS code of the different forwarding layer network includes the IFS code of the first layer network, the IFS code of the second layer network and the IFS code of the third layer network.

[0056] The specific implementation steps of the step 2 are as follows:

[0057] Step 2.1: generating the IFS code of the first layer network based on the function iterative system;

[0058] Step 2.2: generating the IFS code of the second layer network based on the node information of the first layer network and using the IFS code of the first layer network;

[0059] Step 2.3: generating the IFS code of the third layer network based on the node information of the second layer network and using the IFS code of the second layer network;

[0060] The node information includes a new node, a node edge and a central node.

[0061] The information propagation prediction and tracing method and system based on the iterative function system have the advantages that the information propagation fractal network conforming to the actual situation is constructed by using the iterative function system and the SI infectious disease model, the prediction and tracing of the information propagation are realized, the control of the network public opinion is greatly helpful, and the invention can further aggregate the scattered and potential information, so as to be used for the data analysis of the microblog users or information content by the industry; the system is simple to implement and has high efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0062] Figure 1 Schematic diagram of information propagation fractal network construction;

[0063] Figure 2 IFS code diagram of the first-level forwarding network;

[0064] Figure 3 IFS code diagram of the second-level forwarding network;

[0065] Figure 4 IFS code diagram of the third-level forwarding network;

[0066] Figure 5 Information propagation fractal network diagram;

[0067] Figure 6 Schematic diagram of a two-layer fractal network;

[0068] Figure 7 This is a schematic diagram of the system. Detailed Implementation

[0069] Example 1

[0070] The present invention describes an information propagation prediction and tracing system based on an iterative function system, such as... Figures 1-7 As shown, it includes a data acquisition module 1, an initial network acquisition module 2, a propagation model module 3, an information aggregation module 6, an information propagation prediction module 7, and an information tracing module 8;

[0071] The data acquisition module 1 is used to acquire first-level forwarding data of a certain event in the network platform. The first-level forwarding data includes information data, forwarding user data, and opinion and comment data attached by users when forwarding messages.

[0072] The initial network acquisition module 2 is used to perform noise reduction processing on the data acquired by the data acquisition module 1 and to construct a basic network;

[0073] The transmission model module 3 includes an iterative function system module 4 and an SI infectious disease model module 5;

[0074] The propagation model module 3 is used to generate an information propagation fractal network with fractal characteristics from the basic network constructed by the initial network acquisition module 2.

[0075] The iterative function system module 4 is used to generate IFS codes through the basic network constructed by the initial network acquisition module 2. The IFS codes are composed of IFS codes from different forwarding layer networks.

[0076] The SI infectious disease model module 5 is used to generate the edge rules of nodes in the network through the basic network constructed by the initial network acquisition module 2;

[0077] The information aggregation module 6 is used for aggregating the dispersed and potential information in the network, that is, aggregating the IFS codes of different forwarding layer networks.

[0078] The information propagation prediction module 7 is used for predicting the information propagation path through the edge rules of the nodes in the network.

[0079] The information tracing module 8 is used for tracing the originator of the information through the edge rules of the nodes in the network.

[0080] The information propagation prediction and tracing method based on the iterative function system comprises the following steps:

[0081] Step 1: obtaining the first-level forwarding data of an event in a network platform through the crawler technology in the data acquisition module 1, and establishing a basic network through the initial network acquisition module 2 by using the obtained first-level forwarding data, wherein the basic network is used as an integral network of affine transformation.

[0082] Step 2: projecting the basic network established in step 1 onto the coordinate axis, wherein the coordinate position of the root user node is (0, 0), and respectively generating an information propagation fractal network with fractal characteristics by using the iterative function system module 4 in the propagation model module 3.

[0083] Step 3: the information propagation prediction module 7 uses the propagation dynamics equation of the SI infectious disease model module 5 to depict the propagation process, so as to realize the prediction of the information propagation.

[0084] Step 4: according to the coordinates of the final receiver of the public opinion information, the initial publisher of the public opinion information is inferred reversely by using the IFS iteration principle of the iterative function system module 4.

[0085] Step 5: based on the viewpoint and comment data attached by the user when forwarding the message obtained in step 1, the dispersed and potential information in the network is aggregated by using the iterative function system module 4, so as to finally generate an information fractal graph with complete information; the attitude and viewpoint of the public are analyzed from the viewpoint and comment data attached by the user when forwarding the message, so as to analyze the event in the network platform.

[0086] When the first-level forwarding data obtained in step 1 establishes the basic network through the initial network acquisition module 2, the initial publisher of the public opinion information is used as the center node, the new node is the forwarding user data, and the node edge is the forwarding relationship.

[0087] The specific implementation steps of step 2 are as follows:

[0088] Step 2.1: taking the root user node n0(0, 0) as the initial publisher of the public opinion information to generate the first-level forwarding network; the IFS code affine transformation set of the first layer is {R1 2 : ω1， ω1, ....ω m The probability set corresponding to the affine transformation set is {p}. 11 ,p 12,..... p 1m Starting from n0, successively select probability sets {p} 11, p 12,..... p 1m The corresponding probabilities in} are used to generate the central node c_node1 = {n} of the second layer network. 11 ,n 12 ,....,n 1m ,};Propagate information from node n0 to nodes in c_node1 with probability δ;

[0089] Step 2.2: Using the nodes in c_node1 as the initial publishers of the next round of public opinion information dissemination, perform the following transformations on each node in c_node1 according to the second-layer IFS code affine transformation set {R2}. 2 : ω1, ω1, ....ω m}, select the corresponding probability {p 21, p 22 ,.....p 2m Perform affine transformations to generate the set of center nodes of the subgraph, c_node2 = {c_node}. 2i |i=1,...,m}, where Use c_node1 with probability δ to move c_node 2i The nodes in the network propagate information, and if the propagation is successful, a node connection is established; the IFS code of the second-level forwarding network is generated according to the affine transformation formula.

[0090] Step 2.3: Set c_node 2i The node in the definition is a newly generated node in the new round of infection, for c_node 2i Each node in the set is based on the affine transformation set {R3} of the IFS code in the third layer. 2 : ω1, ω1, ....ω m}, select the corresponding probability {p 31, p 32 ,.....p 3m Perform radial transformations to generate the center node set c_node3 = {c_node} of the subgraph. 3i |i=1,...,m}, where Use c_node2 with probability δ to move c_node 3i Information is transmitted between nodes, and if the transmission is successful, a node connection is established.

[0091] The specific implementation steps of step 2.1 are as follows:

[0092] The IFS code is used to draw the fractal figure, and a network is randomly determined as the whole of the geometric object; the IFS code is calculated according to the determined network, different affine transformations are generated according to different IFS codes, so as to form a set of compression affine transformations {R 2 : ω1, ω1,.... ω n} according to the area proportion of each sub-image, and a probability set {p1, p2,.... pn} corresponding to the affine transformation set is obtained, wherein p i > 0 is the probability corresponding to the transformation, and satisfies

[0093] The affine transformation of the two-dimensional Euclidean space is defined as ω: R 2 → R 2 , (X , Y) is a point in the two-dimensional Euclidean space, and the affine transformation image is (X', Y'), which is written in matrix form as:

[0094]

[0095] The node is brought into the above expression, and the following is obtained:

[0096]

[0097] Wherein, ω represents the 6 parameters a, b, c, d, e, and f, when a figure is divided into N local sub-figures, the compression mapping set ω n and the corresponding probability p n of the whole to the local sub-figures form the IFS code, denoted as {X: (ω n , p n , n = 1, 2, 3... N)}.

[0098] The propagation model module 3 is specifically implemented as:

[0099] The nodes in the network are divided into two states, namely the infected state and the susceptible state, the infected person infects the susceptible person at a transmission rate of β, and S(t) and I(t) represent the proportion of susceptible persons and infected persons at time t, and S(t) + I(t) ≡ 1; once the susceptible person is infected, he cannot be cured, and finally all people are infected as time t passes; the propagation dynamics equation of the propagation model is:

[0100]

[0101] The propagation model is defined as follows:

[0102] The IFS code probability of different layers is defined as p i(i＝1,2,3) , wherein

[0103]

[0104] wherein n represents the number of nodes generated in the first layer, N is the total number of nodes in the network, p1 is the probability corresponding to the affine transformation of the first layer, p2 is the probability corresponding to the affine transformation of the second layer, and p3 is the probability corresponding to the affine transformation of the third layer.

[0105] The specific implementation of the step 3 is as follows:

[0106] The differential equation of the SI infectious disease model module (5) is a formula of Markov property, as shown in the following formula:

[0107]

[0108] The formula describes a random process of the transition from one state to another state, and the process has the property of "no memory", that is, the probability distribution of the next state can only be determined by the current state, and the events in the time sequence in front of it have nothing to do with it. In the SI infectious disease model module 5, it is represented as the infection of the day is only related to the number of disease infections of the previous day. According to the change rule of the propagation dynamics equation, the situation of the edge in the network and the information diffusion are judged, so as to predict the propagation trend of the information.

[0109] The specific implementation of the step 4 is as follows:

[0110] According to the coordinates of the new nodes receiving the current information, the coordinates of the initial publisher of the public opinion information are obtained by backstepping according to the IFS code, so as to realize the tracing of the information propagation.

[0111] The specific implementation of the step 5 is as follows:

[0112] The first-level forwarding data and the information structure generated by the iterative function system IFS code generated by the iterative function system module 4 in steps 2.1, 2.2 and 2.3 generate an information fractal graph with complete information.

[0113] Embodiment 2

[0114] The information propagation prediction and tracing method based on the iterative function system comprises the following steps:

[0115] Step 1: Obtain the first-level forwarding data of an event on a microblog through a crawler technology, wherein the data includes the views and comments data attached when a user forwards a message; and a basic network is established by using the obtained data, and the basic network is used as the whole network of affine transformation;

[0116] Step 2: Project the base network established in step 1 onto the coordinate axis, with the coordinate position of the root user node being (0, 0), and generate new nodes and node edges in the information propagation fractal network by combining the IFS code generated by the iterative function system module 4 and the information propagation model generated by the SI epidemic model module 5;

[0117] Step 3: Use the propagation dynamics equation in the information propagation model to depict the propagation process of information, thereby realizing the prediction of information propagation;

[0118] Step 4: According to the coordinates of the final recipients of the public opinion information, use the IFS code iteration principle to infer the initial publisher of the public opinion information in reverse;

[0119] Step 5: Based on the opinion and comment data attached when the user forwards the message obtained in step 1, aggregate the trivial information in the iterative function system, which is the IFS code of different forwarding layers, and finally generate an information fractal graph with complete information. Analyze the audience's attitude and opinion from these microblog information with user comments, thereby for industry to analyze microblog users or information content.

[0120] Further, when establishing the base network in step 1 using information forwarding data, the initial publisher of the public opinion information is taken as the center node, the forwarding user data is taken as the new node, and the forwarding relationship is taken as the node edge.

[0121] Further, the specific implementation of step 2 is:

[0122] Step 2.1: Take the root node n0(0, 0) as the initial publisher of the initial public opinion information to generate the first level forwarding network. The IFS code affine transformation set of the first layer is {R1 2 : ω1, ω1,.... ω m}, and the probability set corresponding to the affine transformation set is {p 11 , p 12 ,.....p 1m}. From n0, select the corresponding probability from the probability set {p 11, , p 12 ,.....p 1m} in turn to generate the center nodes c_node1 = {n 11 , n 12 ,...., n 1m ,} of the second layer network. Use the node n0 to propagate information to the nodes in c_node1 with a probability δ, and if the propagation is successful, an edge is established, as shown in Fig. Figure 1 (a). The IFS code of the first level forwarding network is shown in Fig. Figure 2 .

[0123] Step 2.2: Using the nodes in c_node1 as the initial publishers of the next round of public opinion information dissemination, perform the following transformations on each node in c_node1 according to the second-layer IFS code affine transformation set {R2}. 2 : ω1, ω1, ....ω m}, select the corresponding probability {p 21, p 22 ,.....p 2m Perform affine transformations to generate the set of center nodes of the subgraph, c_node2 = {c_node}. 2i |i=1,...,m}, where Use c_node1 with probability δ to move c_node 2i The nodes in the network propagate information, and if the propagation is successful, an edge is established, as shown in the appendix. Figure 1 As shown in (b), the IFS code for the second-level forwarding network is generated according to the affine transformation formula. The IFS code for the second-level forwarding network is attached. Figure 3 As shown.

[0124] Step 2.3: Set c_node 2i The node in the definition is a newly generated node in the new round of infection, for c_node 2i Each node in the set is based on the affine transformation set {R3} of the third-layer IFS. 2 : ω1, ω1, ....ω m}, select the corresponding probability {p 31 ,p 32 ,.....p 3m Perform radial transformations to generate the center node set c_node3 = {c_node} of the subgraph. 3i |i=1,...,m}, where Use c_node2 with probability δ to move c_node 3i The nodes in the network propagate information, and if the transmission is successful, a connection is established, as shown in the appendix. Figure 1 As shown in (c). Similarly, the IFS code for the third-level forwarding network is shown in the appendix. Figure 4 As shown.

[0125] The principle of the iterative function system in step 2.1 is as follows:

[0126] The iterative function system draws fractal diagrams using IFS codes, randomly selecting a network as the entire geometric object. IFS codes are calculated based on the selected network, and different IFS codes generate different affine transformations, thus forming a compressed affine transformation set {R}. 2 : ω1, ω1, ....ω n Based on the area ratio of each sub-image, the probability set {p1, p2, ..., p} corresponding to the affine transformation set is obtained.n}, where p i >0 is the probability corresponding to the transformation, and satisfies

[0127] Define the affine transformation of two-dimensional Euclidean space as ω:R 2 →R 2 ,(X , Let Y be a point in two-dimensional Euclidean space, and let its affine transformation image be (X', Y'), which can be written in matrix form as:

[0128]

[0129] Substituting the nodes into the above expression, we get:

[0130]

[0131] Here, ω represents the six parameters a, b, c, d, e, and f. When a graph is divided into N local subgraphs, the set of compressed mappings from the whole to the local is ω. n and its corresponding probability p n The IFS code is formed and denoted as {X:(ω n ,p n ,n=1,2,3...N)}.

[0132] Furthermore, the information propagation model in step 2 is specifically implemented as follows:

[0133] This invention constructs a new transmission model based on the classic SI infectious disease model. In the classic SI infectious disease model, nodes in a network are divided into two states: infected and susceptible. Infected individuals infect susceptible individuals with a transmission rate of β. Let S(t) and I(t) represent the ratio of susceptible to infected individuals at time t, and we have S(t) + I(t) ≡ 1. Once infected, susceptible individuals cannot be cured, and eventually, as time t progresses, everyone becomes infected. The transmission dynamics equation characterizing this model is:

[0134]

[0135] The information propagation model in step 2 is defined as follows:

[0136] Define the probability of selecting different layers of IFS as p i(i＝1,2,3) ,in

[0137]

[0138] in n represents the number of nodes generated in the first layer.

[0139] Furthermore, step 3 is implemented as follows:

[0140] The differential equation of the above information propagation model is rewritten into a formula with Markov property, as shown in the following formula:

[0141]

[0142] The above formula describes a random process of state transition to another state, which has the property of "no memory", that is, the probability distribution of the next state can only be determined by the current state, and the events before it in the time sequence are irrelevant. In the infectious disease model, it is expressed as the infection of the day is only related to the number of people infected with the disease the day before. According to the change rule of the propagation dynamics equation, the situation of the edge in the network and the information diffusion are judged, so as to predict the trend of information propagation, as shown in the accompanying Figure 5

[0143] Further, the specific implementation of step 4 is:

[0144] According to the coordinates of the current information receiving node, the coordinates of the information sender are obtained by backtracking according to the IFS code, so as to realize the tracing of the information propagation.

[0145] Further, the specific implementation of step 5 is:

[0146] The IFS code is generated by the iterative function system in steps 2.1, 2.2 and 2.3. In the process of forwarding information on microblog, users add their own opinions and comments to the original information, but the complete structure of the information is generated according to the information content data and the IFS code iteration obtained in step 1. Therefore, in the process of information propagation, the information structure forms an information fractal graph, as shown in the accompanying Figure 6 (b).

[0147] Example 3

[0148] The information propagation prediction and tracing method based on the iterative function system comprises the following steps: the system based on which is the same as that of example 1.

[0149] Step 1: Obtain the first-level forwarding data of a certain event in microblog and the user data of the forwarding message, which is used to establish a basic network; at the same time, obtain the opinion and comment data added by the user when forwarding the microblog;

[0150] Step 2: Based on the basic network obtained in step 1, generate an information propagation fractal network by using the iterative function system and the SI infectious disease model;

[0151] Step 3: The propagation dynamics equation in the SI infectious disease model is used to describe the propagation process of information, so as to realize the prediction of information propagation;

[0152] ​Step 4: The initial information publisher coordinates are calculated reversely by using IFS code to realize accurate positioning of the opinion information publisher, and to control the spread of opinion information from the source;

[0153] Step 5: Based on the content data of the information obtained in step 1, the fractal graph of the information content is generated by using the iterative function system to realize the aggregation of trivial information.

[0154] Further, the specific implementation of step 2 is:

[0155] Step 2.1: The IFS code of the first-level forwarding network is generated by using the iterative function system, and the nodes of the first-level forwarding network are determined according to the IFS code; based on the classical SI infectious disease model, a new information propagation model is established, and the node edges between nodes are established through propagation;

[0156] Step 2.2: Similarly, the IFS code of the second-level forwarding network is generated by using the iterative function system, and the node edges of the nodes in the second-level forwarding network are constructed according to the information propagation model;

[0157] Step 2.3: Similarly, the IFS code of the third-level forwarding network is generated by using the iterative function system, and the node edges of the nodes in the third-level forwarding network are constructed according to the information propagation model.

[0158] Further, in step 2.1, the initial information publisher user is regarded as the initial "infector", and the new nodes generated based on the iterative function system are "infected" according to a certain propagation probability by using the information propagation model, and the edges are established if "infected".

[0159] Further, in step 3, the Markov property of the propagation dynamics equation is used to iteratively analyze the diffusion of information.

[0160] Further, in step 4, the root node information is calculated reversely according to the coordinate information of the current opinion information receiver by using the IFS code, so as to realize the tracing of the information.

[0161] Further, in step 5, the rules for generating IFS code in steps 2.1, 2.2 and 2.3 are used to generate the information fractal graph based on the process of forwarding information by users.

Claims

1. A method for information propagation prediction and source tracing based on iterative function systems, characterized in that: Includes the following steps: Step 1: Obtain first-level forwarding data of a certain event in the network platform through the crawler technology in the data acquisition module (1), and establish the basic network through the initial network acquisition module (2) using the obtained first-level forwarding data. The basic network serves as the overall network of affine transformation. Step 2: Project the basic network established in Step 1 onto the coordinate axes, where the coordinate position of the root user node is (0, 0). Use the iterative function system module (4) in the propagation model module (3) to generate an information propagation fractal network with fractal characteristics. The specific implementation steps of Step 2 are as follows: Step 2.1: Set the root user node As the initial publisher of public opinion information, it generates the first-level forwarding network; The first layer of IFS code affine transformation set is The probability set corresponding to the affine transformation set is ;from Starting from there, select probability sets in sequence. Based on the corresponding probabilities, the central nodes of the second-layer network are generated. ; using nodes With probability Towards Information is propagated from nodes within the system; Step 2.2: [The text appears to be incomplete and contains several grammatical errors. A more accurate translation would require the full context.] The nodes in the chain act as the initial publishers of public opinion information in the next round of dissemination. Each node in the set is based on the affine transformation set of the second-layer IFS code. Choose the corresponding probability Perform affine transformations to generate the set of center nodes for each subgraph. ,in ;use With probability Towards The nodes in the network propagate information, and if the propagation is successful, a node connection is established; the IFS code of the second-level forwarding network is generated according to the affine transformation formula. Step 2.3: [The text appears to be incomplete and contains several grammatical errors. A more accurate translation would require the full context.] The nodes in the definition are newly generated nodes in a new round of infection. Each node in the set is based on the affine transformation set of the IFS code in the third layer. Choose the corresponding probability Perform radial transformations to generate the center node sets of each subgraph. ,in ;use With probability Towards Information is transmitted between nodes, and if the transmission is successful, a node connection is established. Step 3: The information dissemination prediction module (7) uses the dissemination dynamics equation of the SI infectious disease model module (5) to characterize the dissemination process and realize the prediction of information dissemination; Step 4: Based on the coordinates of the final recipient of the public opinion information, use the IFS iteration principle of the iterative function system module (4) to reversely infer the initial publisher of the public opinion information; Step 5: Based on the opinions and comments data attached to the user's forwarded message obtained in Step 1, the scattered and potential information in the network is aggregated using the iterative function system module (4), and finally an information fractal graph with complete information is generated; the public's attitudes and opinions are analyzed from the opinions and comments data attached to the user's forwarded message, so as to analyze a certain event in the network platform.

2. The information propagation prediction and source tracing method based on an iterative function system as described in claim 1, characterized in that: When the primary forwarding data obtained in step 1 establishes the basic network through the initial network acquisition module (2), the initial publisher of public opinion information is taken as the central node, the new node is the forwarding user data, and the node connection is the forwarding relationship.

3. The information propagation prediction and source tracing method based on an iterative function system as described in claim 1, characterized in that: The specific implementation steps of step 2.1 are as follows: The iterative function system plots fractal diagrams using IFS codes. A network is randomly selected as the entire geometric object. IFS codes are calculated based on the selected network, and different affine transformations are generated based on different IFS codes, thus forming a compressed affine transformation set. Based on the area ratio of each sub-image, the probability set corresponding to the affine transformation set is obtained. ,in It is the probability corresponding to the transformation, and satisfies ; Define the affine transformation of two-dimensional Euclidean space as , Let be a point in two-dimensional Euclidean space, and its affine transformation image be... Written in matrix form: ； Substituting the nodes into the above expression, we get: ； in, That means These 6 parameters, when a graphic is divided into When dealing with local subgraphs, the set of compressed mappings from the whole to the local is... and its corresponding probability The IFS code is formed and denoted as .

4. The information propagation prediction and source tracing method based on an iterative function system as described in claim 1, characterized in that: The propagation model module (3) is specifically implemented as follows: Nodes in a network are divided into two states: infected and susceptible. Infected nodes are classified as follows: The transmission rate of infection to susceptible individuals, using express The ratio of susceptible individuals to infected individuals at any given time has Susceptible individuals, once infected, cannot be cured and will eventually succumb to the disease over time. As time progresses, everyone becomes infected; the transmission dynamics equation of the transmission model is: ； The propagation model is defined as follows: Define the probability of selecting IFS codes of different layers as follows: ,in ; ； Where α is the initial infection rate of the information; , This indicates the number of nodes generated in the first layer. This represents the total number of nodes in the network. Let be the probability corresponding to the first layer of affine transformation. This represents the probability corresponding to the second layer of affine transformation. This represents the probability corresponding to the third layer of affine transformation.

5. The information propagation prediction and source tracing method based on an iterative function system as described in claim 1, characterized in that: The specific implementation of step 3 is as follows: The differential equation of the SI infectious disease model module (5) is a formula with Markov properties, as shown in the following equation: ； The formula describes a stochastic process of transition from one state to another. This process has the property of "memorylessness", that is, the probability distribution of the next state can only be determined by the current state. In the time series, the events before it are irrelevant to it. In the SI infectious disease model module (5), it is represented that the infection situation on the day is only related to the number of people infected with the disease on the previous day. According to the change law of the propagation dynamics equation, the connection situation and information diffusion situation in the network are judged, thereby predicting the propagation trend of information.

6. The information propagation prediction and source tracing method based on an iterative function system as described in claim 1, characterized in that: The specific implementation of step 4 is as follows: Based on the coordinates of the current information receiving node, the coordinates of the initial publisher of the public opinion information can be derived by using the IFS code, thereby enabling the tracing of the source of information dissemination.

7. The information propagation prediction and source tracing method based on an iterative function system as described in claim 1, characterized in that: The specific implementation of step 5 is as follows: The first-level forwarding data and the IFS code generated by the iterative function system module (4) in steps 2.1, 2.2, and 2.3 are used to iteratively generate an information fractal network with complete information.

8. The information propagation prediction and source tracing method based on an iterative function system as described in claim 1, characterized in that: IFS codes for different forwarding layer networks include IFS codes for the first layer network, IFS codes for the second layer network, and IFS codes for the third layer network. The specific implementation steps of step 2 are as follows: Step 2.1: Generate the IFS code of the first layer network based on the function iterative system; Step 2.2: Based on the node information of the first-layer network, generate the IFS code of the second-layer network using the IFS code of the first-layer network; Step 2.3: Based on the node information of the second-layer network, generate the IFS code of the third-layer network using the IFS code of the second-layer network; The node information includes newly created nodes, node connections, and the central node.

9. A system for performing an information propagation prediction and tracing method based on an iterative function system as described in claim 1, characterized in that: It includes a data acquisition module (1), an initial network acquisition module (2), a propagation model module (3), an information aggregation module (6), an information propagation prediction module (7), and an information tracing module (8). The data acquisition module (1) is used to acquire first-level forwarding data of a certain event in the network platform. The first-level forwarding data includes information data, forwarding user data, and opinion and comment data attached when the user forwards the message. The initial network acquisition module (2) is used to perform noise reduction processing on the data acquired by the data acquisition module (1) and to construct a basic network; The propagation model module (3) includes an iterative function system module (4) and an SI infectious disease model module (5). The propagation model module (3) is used to generate an information propagation fractal network with fractal characteristics from the basic network constructed by the initial network acquisition module (2). The iterative function system module (4) is used to generate IFS codes through the basic network constructed by the initial network acquisition module (2). The IFS codes are composed of IFS codes from different forwarding layer networks. The SI infectious disease model module (5) is used to generate the connection rules of nodes in the network through the basic network constructed by the initial network acquisition module (2); The information aggregation module (6) is used to aggregate scattered and potential information in the network, that is, to aggregate IFS codes of different forwarding layer networks. The information propagation prediction module (7) is used to predict the information propagation path through the connection rules of nodes in the network; The information tracing module (8) is used to reverse the initiator of information by using the connection rules of nodes in the network, so as to realize the tracing of information.

Images