CN112800336B - Online social network user behavior prediction method based on simple harmonic vibration theory - Google Patents

Online social network user behavior prediction method based on simple harmonic vibration theory Download PDF

Info

Publication number
CN112800336B
CN112800336B CN202110167680.6A CN202110167680A CN112800336B CN 112800336 B CN112800336 B CN 112800336B CN 202110167680 A CN202110167680 A CN 202110167680A CN 112800336 B CN112800336 B CN 112800336B
Authority
CN
China
Prior art keywords
user
information
influence
network topology
network
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Fee Related
Application number
CN202110167680.6A
Other languages
Chinese (zh)
Other versions
CN112800336A (en
Inventor
罗文浩
谭振华
刘春晓
孙治强
鲁钰娟
赵诗淇
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Northeastern University China
Original Assignee
Northeastern University China
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Northeastern University China filed Critical Northeastern University China
Priority to CN202110167680.6A priority Critical patent/CN112800336B/en
Publication of CN112800336A publication Critical patent/CN112800336A/en
Application granted granted Critical
Publication of CN112800336B publication Critical patent/CN112800336B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9535—Search customisation based on user profiles and personalisation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9536—Search customisation based on social or collaborative filtering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/01—Social networking

Landscapes

  • Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Economics (AREA)
  • Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Human Resources & Organizations (AREA)
  • Marketing (AREA)
  • Primary Health Care (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

The invention discloses an online social network user behavior prediction method based on a simple harmonic vibration theory, which comprises the following steps of: splitting each user data in the online social network space into information which is released by a user history and a network topology structure where the user is located; respectively calculating the influence of the information historically issued by each user and the network topology influence of the network topology structure of the user; obtaining the comprehensive influence of each user according to the influence of information historically issued by the user of each user and the influence of the network topology structure where the user is located; and setting a threshold value to predict the decision behavior of whether each user can forward the information under the action of the comprehensive influence of the corresponding user. The method can be used for monitoring network public sentiments, can predict the general trend of rumors at the initial period of rumor explosion, and accurately positions the future network behavior of a certain user.

Description

Online social network user behavior prediction method based on simple harmonic vibration theory
Technical Field
The invention relates to the field of network space safety, in particular to an online social network user behavior prediction method based on a simple harmonic vibration theory.
Background
The research on the information propagation behavior in the online social network space is used for explaining how information is transmitted among network users, and because each user is an independent individual in real life, the behavior of the whole netizen group is extremely complex, thereby bringing great difficulty to the research work in the field. In the field of online social network space security, a category of information propagation behaviors which can bring great instability factors to the society is rumor information propagation, if the propagation rule of the rumor information can be grasped in time, the future behavior of a certain netizen can be accurately predicted, and the rumor can be prevented from being diffused in a large scale by adopting a rumor restraining means in advance. However, most of the existing technical schemes throughout the year only have the estimation of the final influence scale after the rumor occurs, and can only passively process the retrospective work of the adverse influence caused after the rumor occurs, and the prevention cannot be realized.
The research on information dissemination behaviors in online social networks is long, but few researches emphasize the role of information contents per se and only research on information dissemination rules through the attention relationship among demographics. However, many current technologies in this field only have approximate prediction on macroscopic trends, cannot achieve microscopic depiction of future behaviors of a certain user, and most existing technologies are designed based on a classical infectious disease model, but information transmission and disease infection on an online social network are not completely equivalent.
Disclosure of Invention
The invention aims to solve the defects in the prior art and provides an online social network user behavior prediction method based on a simple harmonic vibration theory.
In order to achieve the purpose, the invention is implemented according to the following technical scheme:
an online social network user behavior prediction method based on a simple harmonic vibration theory comprises the following steps:
s1, splitting each user data in the online social network space into information published by the user history and a network topology structure where the user is located;
s2, respectively calculating the influence of the information historically issued by each user and the influence of the network topology structure where the user is located;
s3, acquiring the comprehensive influence of each user according to the influence of the information historically issued by each user and the network topology influence of the network topology structure where the user is located;
and S4, setting a threshold value to predict the decision behavior of whether each user will forward the information under the comprehensive influence of the corresponding user.
Further, the specific step of calculating the influence of the information historically issued by each user in step S2 is as follows:
s21, recording the information content of all user history releases in a certain fixed community on a social network site, filtering and segmenting each piece of information of the user history releases, removing stay words to generate a training set, inputting the training set into a Word2vector neural network for training to obtain a Word vector model, outputting a plurality of Word vectors of each piece of information of the user history releases through the Word vector model, and adding corresponding position elements of the Word vectors on the space to obtain the information vector of each piece of information of the user history releases:
Figure BDA0002937975150000021
in the formula: vocvec1，vocvec2…vocvecnA word vector representing n 300 dimensions;
s22, obtaining the interest vector of the user at the current time point through the information vector of each piece of information released by each user history:
Figure BDA0002937975150000031
in the formula: m represents the quantity of information historically published by each user;
s23, using cosine similarity between the information vector about to happen at the current time and the interest vector of the user at the current time point as the influence p of the information on the information historically issued by the userinfo2vFormalized as follows:
Figure BDA0002937975150000032
in the formula:
Figure BDA0002937975150000033
is the information vector and the 2-norm,
Figure BDA0002937975150000034
and
Figure BDA0002937975150000035
are the user interest vector and the 2 norm.
Further, the specific step of calculating the network topology influence of the network topology structure where the user of each user is located in step S2 is as follows:
s24, defining the number of users concerned by the user and fan users in the network topology structure of the user as the input degree number and the output degree number;
s25, establishing and measuring network topology energy of the user in the network topology structure of the user based on the out-degree number of the user node:
Figure BDA0002937975150000036
wherein:
Figure BDA0002937975150000037
lg represents the logarithm of the common base number of 10, AiEnergy possessed by information publishing users in a network topology structure in which the users with the depth of 1 layer are located, AiThe value of the node represents the strength of the information output capability of the current node in the network topology structure where the user is located;
s26, the influence transmitted by any user is transmitted to the fan users of the users in the form of simple harmonic waves, namely, the fluctuation curve of the influence
Figure BDA0002937975150000038
Ai>0，AiThe network topology energy output to the target node by the source node is shown, and omega is the initial circular frequency of the network;
Figure BDA0002937975150000041
is the initial phase of the simple harmonic wave, define
Figure BDA0002937975150000042
Abstracting a degree of behavioral synchronization between users regarding relationships in a social network;
defining degrees of behavioral similarity between users regarding annotation relations:
Figure BDA0002937975150000043
in the formula: i SvI is the number of all behavior records in the user history, Countu&vThe quantity of the same information forwarded among the users with the directed relationship; when S isvWhen the value is 0, the degree of similarity of behaviors between users related to the annotation relation is 0.01;
the degree of behavioral synchronization between users regarding the annotation relationship
Figure BDA0002937975150000044
The modeling is defined as follows:
Figure BDA0002937975150000045
s27, sending a topological influence fluctuation to the user in a simple harmonic mode by a plurality of users having a directed relationship with the user, and obtaining n equivalent simple harmonics after the simple harmonic fluctuation is superposed according to a simple harmonic superposition theory as follows:
Figure BDA0002937975150000046
in the equivalent simple harmonic expression
Figure BDA0002937975150000047
Wherein A is the amplitude of the new waveform obtained by superposing the waveforms of the rows,
Figure BDA0002937975150000048
is the equivalent initial phase of the new waveform;
s28, the mathematical description of the network topology size from the source node obtained by correcting the equivalent topology energy is as follows:
Figure BDA0002937975150000049
s29, processing the network topology energy by using a Sigmod function to constrain the network topology energy to be 0-1, and obtaining the network topology influence of the network topology structure on the target node:
Figure BDA0002937975150000051
in the formula: p is a radical ofuvNamely the network topology influence of the network topology structure of the user.
Further, the expression of the comprehensive influence of each user in step S3 is:
pv＝puv*Pinfo2v；
in the formula: p is a radical ofuvNetwork topology influence, p, for the network topology of each userinfo2vThe influence of information published for each user's user history.
Further, the step S4 specifically includes:
setting a threshold value T, when the user's integrated influence pvWhen the new information is more than or equal to T, predicting that the user is most likely to participate in the forwarding of the new information in the future; otherwise, the probability of the user participating in the information forwarding in the future is predicted to be extremely low.
The invention provides an online social network user behavior prediction method based on a simple harmonic vibration theory, which can be used for monitoring network public sentiment, can predict the general trend of rumors in the initial period of rumor outbreak, and accurately positions the future network behavior of a certain user. Compared with the existing mature technology, the invention has the advantages of three aspects:
(1) through the high similarity of the fields of cyberspace information propagation and mechanical vibration, the propagation of influence in the online social network is compared with the propagation of simple harmonic fluctuation, which is an attempt never made in the online social network space.
(2) The influence of a social network topological structure and information content on the user is fused, wherein the information content part constructs the personalized preference of the network user to the information by constructing the similarity of the information vector and the user interest vector, and the characteristics of the network user can be more carefully described.
(3) The method can accurately predict the action of whether the user will forward the information in the future in the online social network on both a macroscopic level and a microscopic level.
Drawings
FIG. 1 is a flow chart of the present invention.
Fig. 2 is a detailed diagram of the technical implementation of the invention.
Fig. 3 is a word vector visualization diagram of three words from the surf microblog data in the embodiment of the invention.
Fig. 4 is a comparison graph of macroscopic trends of 6 real new wave microblog information diffusion data sets in an application example.
Fig. 5 is a macroscopic instantaneous increment comparison diagram of 6 real surf microblog information diffusion data sets in an application example.
FIG. 6 is a visualization of the true diffusion process of the propagation process of information 3(Info3) in the application example: (a) time step 1 hop; (b) time step 2 hop; (c) time step 5 hop.
FIG. 7 is a visualization of the diffusion process of information 3(Info3) in the application example, which is propagated under the method of the present invention: (a) time step 1 hop; (b) time step 2 hop; (c) time step 5 hop.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. The specific embodiments described herein are merely illustrative of the invention and do not limit the invention.
As shown in fig. 1 and fig. 2, the present embodiment provides an online social network user behavior prediction method based on simple harmonic vibration theory, including the following steps:
the method comprises the steps that firstly, each user data in an online social network space is divided into information published by a user history and a network topological structure where the user is located; the information published by the user contains the interest and the preference of the user, and the topological structure contains the channel for receiving the information by the user.
Secondly, respectively calculating the influence of the information historically issued by each user and the network topology influence of the network topology structure where the user is located; information content is a carrier of information influence propagation in online social networks, where the influence of information content on users is the driving force of the essence of information diffusion. In this embodiment, the interests and information contents of the user are constructed in a vector form, and the similarity between the information contents and the interests of the user is used as the influence of the information contents on the user.
Recording the information content of all user historical releases in a certain fixed community on a social network site, filtering and segmenting each piece of information of the user historical releases, generating a training set after removing stay words, inputting the training set into a Word2vector neural network for training to obtain a Word vector model, and outputting a plurality of Word vectors of each piece of information of the user historical releases through the Word vector model. In order to verify the feasibility of vectorization of words, in this embodiment, three words are selected from the dataset of the Xinlang microblog, which are voc1 ═ JayChou ', voc2 ═ SelbyAbbey ', and voc3 ═ Dior ', respectively. Where voc1 is a famous singer in china, voc2 is a famous council in the uk, and voc3 is a famous luxury brand. We found that there is an association between voc1 and voc2, and that there is no association between voc3 and the former two. The word vectors with the length of 300 dimensions obtained by the word vector model are respectively as follows:
vector1＝[-0.00398873，0.00338546，0.02178973....]300；
vector2＝[-0.0050188，0.01185059，0.02922338....]300；
vector3＝[-0.23428404，0.39206254，0.38281527....]300；
the three vectors are visualized for visual observation of the differences between the three vectors, and the results are shown in fig. 3.
From the word vector comparison of 'JayChou' and 'SelbyAbbey' in fig. 3, it is found that the two maps are very similar, and on the contrary, the map of 'color' is greatly different from the first two maps, which is consistent with the real information content relevance, so that the vectorization of the vocabulary can well distinguish the information contents with different relevance degrees.
Suppose that n 300-dimensional word vectors voc are obtained after word segmentation processing and vectorization of certain microblog informationvec1，vocvec2…vocvecnAdding the corresponding position elements of the n word vectors on the space to obtain the information vector info of the informationvec：
Figure BDA0002937975150000071
In the formula: vocvec1，vocvec2…vocvecnA word vector representing n 300 dimensions;
thus infovecAlso a vector of dimension 300. Then, assuming that the current user publishes m pieces of information in total, the interest vector of the user at the current time point is obtained through the information vector of each piece of information published by each user history:
Figure BDA0002937975150000081
in the formula: m represents the quantity of information historically published by each user;
using cosine similarity between the information vector about to occur at the current time and the interest vector of the user at the current time point as the influence p of the information on the information historically issued by the userinfo2vFormalized as follows:
Figure BDA0002937975150000082
in the formula:
Figure BDA0002937975150000083
is a direction of informationThe amount and the 2-norm,
Figure BDA0002937975150000084
and
Figure BDA0002937975150000085
are the user interest vector and the 2 norm.
An important feature in the field of online social network research, in which we often refer to the number of 2 as degree-in and degree-out, is the topological structure of the network, determined by the users who are interested in the user and the fan users.
The network topology structure is an 'ecological environment' depending on users in the online social network, and provides a path for information diffusion. An important topological feature in network topology is the degree of the user node. In an online social network, the out-degree number represents the number of channels for a user to transmit information, and the in-degree number represents the number of channels for a user to receive information transmitted by others. In the information diffusion research, the degree of the user node plays a significant role in the construction of the user influence.
Therefore, the number of concerned users and fan users in the network topology structure where the users are located is defined as the input degree number and the output degree number; establishing and measuring network topology energy of the user in a network topology structure of the user based on the out-degree number of the user node:
Figure BDA0002937975150000091
wherein:
Figure BDA0002937975150000092
lg represents the logarithm of the common base number of 10, AiEnergy possessed by information publishing users in a network topology structure in which the users with the depth of 1 layer are located, AiThe value of the node represents the strength of the information output capability of the current node in the network topology structure where the user is located;
due to any one userThe transmitted influence is transmitted to fan users of users in the form of simple harmonic waves, namely an influence fluctuation curve
Figure BDA0002937975150000093
Ai>0，AiThe network topology energy output to the target node by the source node is shown, and omega is the initial circular frequency of the network;
Figure BDA0002937975150000094
is the initial phase of the simple harmonic wave, define
Figure BDA0002937975150000095
Abstracting a degree of behavioral synchronization between users regarding relationships in a social network;
defining degrees of behavioral similarity between users regarding annotation relations:
Figure BDA0002937975150000096
in the formula: i SvI is the number of all behavior records in the user history, Countu&vThe quantity of the same information forwarded among the users with the directed relationship; when S isvWhen the value is 0, the degree of similarity of the behaviors between the users related to the annotation relation is 0.01, and the degree of synchronization of the behaviors between the users related to the annotation relation
Figure BDA0002937975150000097
The modeling is defined as follows:
Figure BDA0002937975150000098
a plurality of users having a directed relationship with the users can send a topological influence fluctuation to the users in a simple harmonic mode, and n equivalent simple harmonics after the simple harmonic fluctuation is superposed are obtained according to a simple harmonic superposition theory as follows:
Figure BDA0002937975150000099
in the equivalent simple harmonic expression
Figure BDA0002937975150000101
Wherein A is the amplitude of the new waveform obtained by superposing the waveforms of the rows,
Figure BDA0002937975150000102
is the equivalent initial phase of the new waveform; the mathematical description of the network topology size from the source node obtained by correcting the equivalent topology energy is as follows:
Figure BDA0002937975150000103
processing the network topology energy by using a Sigmod function to constrain the network topology energy to be between 0 and 1, and obtaining the network topology influence of a network topology structure on a target node:
Figure BDA0002937975150000104
in the formula: p is a radical ofuvNamely the network topology influence of the network topology structure of the user.
A third step whereby the power p is influenced in the network topologyuvAnd information content influence pinfo2vUnder the common influence of the two, obtaining the comprehensive influence of each user according to the influence of the information historically issued by each user and the influence of the network topology structure where the user is located:
pv＝puv*Pinfo2v；
in the formula: p is a radical ofuvNetwork topology influence, p, for the network topology of each userinfo2vThe influence of information published for each user's user history.
And fourthly, setting a threshold value to predict the decision behavior of whether each user can forward the information under the action of the comprehensive influence of the corresponding user.
Setting a threshold value T, when the user's integrated influence pvWhen T is more than or equal to T, predicting that the user is most likely to participate in the forwarding of the new information in the future, wherein the state S is changed into I in FIG. 2; otherwise, the probability of the user participating in the information forwarding in the future is predicted to be extremely low.
Examples of the applications
In order to verify the feasibility and reliability of the technology, 6 information forwarding processes are selected from the social network of the Xinlang microblog, and the propagation process of the 6 information is described as shown in table 1.
TABLE 1
Depth Info1 Info2 Info3 Info4 Info5 Info6
1(seed) 1 1 1 1 1 1
2 445 364 394 554 621 378
3 266 186 285 219 98 155
4 41 17 86 10 103 150
5 2 7 36 3 50 46
6 0 0 0 0 5 17
7 0 0 0 0 0 0
Total up to 755 575 802 787 878 747
By applying the method of the invention to the 6 real Sing microblog information diffusion data sets respectively, fitting comparison graphs of macroscopically and real propagation processes are obtained as shown in the following fig. 4 and 5, wherein a solid line represents the real propagation process, and a dotted line represents a prediction result of the novel technology invented by the inventor. Experimental results show that the technology provided by the inventor can well predict the propagation trend of certain information in the macroscopic aspect.
In addition, we have selected a visual comparison graph in which the propagation process of the information 3(Info3) is respectively performed with the real result and the result of the method of the present invention, as shown in fig. 6 and fig. 7, respectively, where the lighter dots represent netizens who do not participate in the information forwarding at the current time, the darker dots represent netizens who have just participated in the information forwarding at the current time, and we have selected two processes to be plotted corresponding to the result at the same time.
On the microscopic picture of user behavior, an independent cascade model commonly used in the field of online social network space security is selected for carrying out accuracy experiment comparison, an accuracy performance index is used, the formula is as follows,
Figure BDA0002937975150000111
when the information transmission simulated by the method of the invention is finished, all netizens participating in information transmission in the network are also in the proportion participating in the transmission in the real transmission process. The results of these 2 technical comparisons are shown in table 2.
TABLE 2
PRE IC SHW
Info1 0.580715 0.794549
Info2 0.767606 0.858770
Info3 0.658965 0.829132
Info4 0.709467 0.846749
Info5 0.674387 0.762264
Info6 0.897314 0.957173
AVG 0.714742 0.841440
As can be seen from Table 2: the method provided by the invention has higher accuracy and can realize microcosmically and accurately predicting the behavior of netizens in the online social network. The online social network rumor control method creates a new idea for behavior prediction of information propagation in the online social network space, and hopefully, new contributions can be made to other safety problems such as rumor control in the online social network.
The technical solution of the present invention is not limited to the limitations of the above specific embodiments, and all technical modifications made according to the technical solution of the present invention fall within the protection scope of the present invention.

Claims (3)

1. An online social network user behavior prediction method based on a simple harmonic vibration theory is characterized by comprising the following steps:
s1, splitting each user data in the online social network space into information published by the user history and a network topology structure where the user is located;
s2, respectively calculating the influence of the information historically issued by each user and the influence of the network topology structure where the user is located:
s21, recording the information content of all user history releases in a certain fixed community on a social network site, filtering and segmenting each piece of information of the user history releases, removing stay words to generate a training set, inputting the training set into a Word2vector neural network for training to obtain a Word vector model, outputting a plurality of Word vectors of each piece of information of the user history releases through the Word vector model, and adding corresponding position elements of the Word vectors on the space to obtain the information vector of each piece of information of the user history releases:
Figure FDA0003522513660000011
in the formula: vocvec1，vocvec2…vocvecnA word vector representing n 300 dimensions;
s22, obtaining the interest vector of the user at the current time point through the information vector of each piece of information released by each user history:
Figure FDA0003522513660000012
in the formula: m represents the quantity of information historically published by each user;
s23, using cosine similarity between the information vector about to happen at the current time and the interest vector of the user at the current time point as the influence p of the information on the information historically issued by the userinfo2vFormalized as follows:
Figure FDA0003522513660000021
in the formula:
Figure FDA0003522513660000022
is the information vector and the 2-norm,
Figure FDA0003522513660000023
and
Figure FDA0003522513660000024
are the user interest vector and the 2 norm;
s24, defining the number of users concerned by the user and fan users in the network topology structure of the user as the input degree number and the output degree number;
s25, establishing and measuring network topology energy of the user in the network topology structure of the user based on the out-degree number of the user node:
Figure FDA0003522513660000025
wherein:
Figure FDA0003522513660000026
lg represents the logarithm of the common base number of 10, AiEnergy possessed by information publishing users in a network topology structure in which the users with the depth of 1 layer are located, AiThe value of the node represents the strength of the information output capability of the current node in the network topology structure where the user is located;
s26, the influence transmitted by any user is transmitted to the fan user of the user in the form of simple harmonic waves, namely the fluctuation curve of the influence
Figure FDA0003522513660000027
Ai＞0，AiThe network topology energy output to the target node by the source node is shown, and omega is the initial circular frequency of the network;
Figure FDA0003522513660000028
is the initial phase of the simple harmonic wave, define
Figure FDA0003522513660000029
Abstracting a degree of behavioral synchronization between users regarding the annotation relation in the social network;
defining the degree of behavioral similarity between users regarding annotation relationships:
Figure FDA00035225136600000210
in the formula: i SvI is the number of all behavior records in the user history, Countu&vThe quantity of the same information forwarded among the users with the directed relationship; when S isvWhen the value is 0, the degree of similarity of behaviors between users related to the annotation relation is 0.01;
the degree of behavioral synchronization between users regarding the annotation relationship
Figure FDA0003522513660000031
The modeling is defined as follows:
Figure FDA0003522513660000032
s27, sending a topological influence fluctuation to the user in a simple harmonic form by a plurality of users having a directed relation with the user, and obtaining n equivalent simple harmonic waves after the simple harmonic fluctuations are superposed according to a simple harmonic superposition theory as follows:
Figure FDA0003522513660000033
in the equivalent simple harmonic expression
Figure FDA0003522513660000034
Wherein A is the amplitude of the new waveform obtained by superposing the waveforms of the rows,
Figure FDA0003522513660000035
is the equivalent initial phase of the new waveform;
s28, the mathematical description of the network topology size from the source node obtained by correcting the equivalent topology energy is as follows:
Figure FDA0003522513660000036
s29, processing the network topology energy by using a Sigmod function to constrain the network topology energy to be 0-1, and obtaining the network topology influence of the network topology structure on the target node:
Figure FDA0003522513660000037
in the formula: p is a radical ofuvThe network topology influence of the network topology structure of the user is obtained;
s3, obtaining the comprehensive influence of each user according to the influence of the information historically issued by each user and the influence of the network topology structure where the user is located;
and S4, setting a threshold value to predict the decision behavior of whether each user will forward the information under the comprehensive influence of the corresponding user.
2. The method for predicting the user behavior in the online social network based on the simple harmonic vibration theory as claimed in claim 1, wherein the expression of the comprehensive influence of each user in the step S3 is as follows:
pv＝puv*pinfo2v；
in the formula: p is a radical ofuvNetwork topology influence, p, for the network topology of each userinfo2vThe influence of information published for each user's user history.
3. The method for predicting user behavior in an online social network based on the simple harmonic vibration theory as claimed in claim 2, wherein the step S4 specifically comprises:
setting a threshold value T, when the user's combined influence pvWhen the new information is more than or equal to T, predicting that the user is most likely to participate in the forwarding of the new information in the future; otherwise, the probability of the user participating in the information forwarding in the future is predicted to be extremely low.
CN202110167680.6A 2021-02-07 2021-02-07 Online social network user behavior prediction method based on simple harmonic vibration theory Expired - Fee Related CN112800336B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110167680.6A CN112800336B (en) 2021-02-07 2021-02-07 Online social network user behavior prediction method based on simple harmonic vibration theory

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110167680.6A CN112800336B (en) 2021-02-07 2021-02-07 Online social network user behavior prediction method based on simple harmonic vibration theory

Publications (2)

Publication Number Publication Date
CN112800336A CN112800336A (en) 2021-05-14
CN112800336B true CN112800336B (en) 2022-06-17

Family

ID=75814671

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110167680.6A Expired - Fee Related CN112800336B (en) 2021-02-07 2021-02-07 Online social network user behavior prediction method based on simple harmonic vibration theory

Country Status (1)

Country Link
CN (1) CN112800336B (en)

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9031888B2 (en) * 2011-08-10 2015-05-12 International Business Machines Corporation Predicting influence in social networks
CN109829504B (en) * 2019-02-14 2022-07-01 重庆邮电大学 Prediction method and system for analyzing user forwarding behavior based on ICS-SVM
CN111241421B (en) * 2020-01-14 2022-07-01 西安电子科技大学 User forwarding behavior prediction method based on social context information

Also Published As

Publication number Publication date
CN112800336A (en) 2021-05-14

Similar Documents

Publication Publication Date Title
Amplayo et al. Cold-start aware user and product attention for sentiment classification
US8923830B2 (en) Method and apparatus for processing messages in a social network
Sakib et al. A collaborative approach toward scientific paper recommendation using citation context
Chatterjee et al. Single document extractive text summarization using genetic algorithms
Theophilo et al. Authorship attribution of social media messages
Ristanti et al. Cosine similarity for title and abstract of economic journal classification
Engle et al. Deterministic symbolic regression with derivative information: General methodology and application to equations of state
Guo et al. Federated learning for personalized humor recognition
CN112800336B (en) Online social network user behavior prediction method based on simple harmonic vibration theory
CN112487110A (en) Overlapped community evolution analysis method and system based on network structure and node content
Kanavos et al. Fuzzy information diffusion in twitter by considering user’s influence
Khadija et al. Enhancing Indonesian customer complaint analysis: LDA topic modelling with BERT embeddings
Song et al. Detecting positive opinion leader group from forum
Kaya et al. Application of data mining methods for pattern recognition in negotiation support systems
Huang et al. Early detection of fake news based on multiple information features
Wai et al. Ontology based web page classification system by using enhanced C4. 5 and Naïve Bayesian classifiers
Foley et al. Matching pairs: Attributing fine-tuned models to their pre-trained large language models
Yavary et al. Rumor detection on Twitter using extracted patterns from conversational tree
Mahmoud et al. Context identification of scientific papers via agent-based model for text mining (ABM-TM)
Nguyen et al. Supervised learning models for social bot detection: literature review and benchmark
Zheng et al. Embeddingtree: Hierarchical exploration of entity features in embedding
Tian et al. Spammer detection on Sina micro-blog
Pati et al. An Empirical Analysis of Similarity based Single Document Summarization
Saputra et al. Aspect Based Sentiment Analysis Using Recurrent Neural Networks (RNN) on Social Media Twitter
Zhang et al. A deep recommendation framework for completely new users in mashup creation

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20220617

CF01 Termination of patent right due to non-payment of annual fee