社会网络中信息流与舆论演变的多层动态模型研究资源-CSDN文库

版权申诉

79 浏览量 2024-11-30 18:52:39 上传评论收藏 1.02MB PDF 举报

资源推荐

资源详情

资源评论

Team # 49436 Page 1 of 22

For office use only

________________

Team Control Number

49436

Problem Chosen

For office use only

________________

2016

MCM/ICM

Summary Sheet

Who Moved My Opinion?

As society’s networks evolve, information becomes a new kind of property because of its

great power to change public opinion. In our paper, we study the flow of information, the

evolution of communication networks and the information’s influence on publics.

At the beginning, we describe the topologies of two types of communications which are

Mass Communication and Interpersonal Communication. By combining these two types of

communications, we construct the single-layer communication network which represents a

certain communication technique. Then we study the six periods classified by the techniques

and construct the multi-layer information networks based on the single-layer network.

Then we define the three attributes of information which are extremeness, entertainment

and newness. By using the concept of Indifference Curve, we determine the inherent value of

information based on its attributes. And we find the news according to the Pareto Principle.

Combining the multi-layer information networks and the inherent value of information,

we construct the dynamic information flow model. This model is used to describe how

information flows within our society’s networks.

In order to validate our models, we find data of a piece of news of BBC Tech and simulate

the information flow in today’s society. We find that the real amount of received nodes are 44

and our model’s result is 43.85. Also, the standard error of received nodes’ amount is 4.63. So,

the predict capacity and the reliability of our model is great.

We define the capacity as the user percentage, and the relations as the relative strength of

networks. By introducing the Life-Cycle Theory, we build a prediction model to predict the

relations and capacities of networks in 2050. We find that there will be 7 communication

techniques at that time and the strongest network is a rising network whose capacity is 0.9611.

Later, we modify our information flow model by appending the opinion of people. Then

we use this model to describe the mechanism of the change of public opinion based on

Threshold Effect. Then we analyze the influence of four main parameters in our model and get

three main conclusions: (1) Stubborn minorities. (2) Powerful media. (3) More homogeneous

more efficient.

Finally, we give the strengths and weaknesses of our models.

Team # 49436 Page 2 of 22

Contents

1. Introduction .................................................................................................. 3

1.1 Problem Background ............................................................................... 3

1.2 Previous Research .................................................................................... 3

1.3 Our Work .................................................................................................. 3

2. Assumptions and Justification ...................................................................... 4

3. Topology of Information Networks .............................................................. 4

3.1 Two Types of Communications ............................................................... 4

3.2 A Slice of Toast: Construct Single-Layer Network .................................. 6

3.3 Slices of Toast: Construct Multi-Layer Networks ................................... 7

4. Model of Inherent Value of Information ...................................................... 8

4.1 Attributes of Information ......................................................................... 9

4.2 The Inherent Value of Information ........................................................... 9

4.3 Pick News out of Information ................................................................ 10

5. The Dynamic Information Flow Model ..................................................... 10

5.1 The Forwarding Rules of Media Nodes ................................................. 11

5.2 The Forwarding Rules of Person Nodes ................................................ 11

6. How a BBC’s News Disseminated: Validate Our Model ........................... 12

6.1 Find Data for Simulation ....................................................................... 12

6.2 Simulation .............................................................................................. 12

6.3 Results .................................................................................................... 13

7. The Evolution of Information Networks .................................................... 13

7.1 Definitions of capacities and relationships of networks ........................ 14

7.2 Emergence of a New Network ............................................................... 14

7.3 The Peak of Different Networks ............................................................ 14

7.4 Networks’ Evolution Based on Life-Cycle Theory ................................ 14

7.5 Prediction of Society’s Information Networks in 2050 ......................... 16

8. Change People’s Minds: An Opinion-Changing Model ............................. 16

8.1 The Opinion of Publics .......................................................................... 17

8.2 The Opinion-Changing Mechanism ....................................................... 17

9. Sensitivity analyses .................................................................................... 18

9.1 More people agree at first, more people agree at last ............................ 18

9.2 Want a bigger influence, go to find a media .......................................... 19

9.3 More Homogeneous, More Efficient ..................................................... 19

9.4 Media Nodes vs. Person Nodes ............................................................. 20

10. Strengths and Weaknesses .......................................................................... 21

10.1 Strengths .............................................................................................. 21

10.2 Weaknesses .......................................................................................... 21

11. References .................................................................................................. 22

Team # 49436 Page 3 of 22

1. Introduction

1.1 Problem Background

Under a certain network situation, the flow and the speed of information are

influenced by two main factors: (1) The inherent value of information. (2) Whether the

information is forwarded by big media. However, the evolution of the society’s

information network never stops, which allows us to study the evolution of the society’s

functions and structure by taking a historical perspective of flow of information.

In the paper, we have been asked to answer the following questions:

 Establish models to describe the flow of information and find what makes

news.

 Test our models’ reliability and prediction capability.

 Predict the relationships and capacities of 2050’s society information network.

 Define people’s opinion and use our models to find how publics’ interest and

opinion can be changed by information.

 Test our model with the four aspects in the flow of information.

1.2 Previous Research

Scholars have been studying the society’s information networks for a long time.

At the beginning, they use the epidemic model to describe the flow of information. In

1998, D.J.Watts and S.H.Strogatz establish WS Small World Model to show the flow

of information and find that the speed of information is fastest in the Small World

Model[1]. In 2002, Staniford S, Paxson V and Weaver N use the Random Constant

Spread Model to study the information flow on the Internet [2]. D.J.Watts and

P.S.Dodds find that the flow of information is easily influenced by the people who are

easily influenced [3]. Also, there have been many experts studying the evolution of

society’s information networks. In these researches, Ellision.N.B’s and Castells M ‘s

work is impressive [4][5].

1.3 Our Work

We firstly build a single-layer information networks. Then we form a multi-layer

networks by combining single-layer networks with the same nodes. After that, we

define the information forward mechanism to allow information to flow in our networks.

Secondly, we use a real information flow’s example to validate our model. We find data

on the Facebook and use our model to simulate it. Thirdly, we explore the principle of

the information networks’ evolution. Then we make a prediction of capacities and

relationships in 2050.Fourthly, we model how people’s opinion be changed. According

to the threshold effect, we give a certain mechanism of people changing their mind.

Finally, we apply our models to analyze how information value, people’s initial opinion,

information source and topology influence spread range and public opinion.

Team # 49436 Page 4 of 22

2. Assumptions and Justification

 The information does not change when it flows within the social information

networks. In the flow of information, the content of information may change

randomly and it is nearly impossible to consider the random changes in the contents

of information.

 Information’s size is the same. In real world, the size of information will influence

the time. However, the size will only change the time and will not influence the

flow mode.

 The population does not change. The population changes the nodes. But it will

not change the way in which information flows in the networks.

 Information’s value satisfies the Pareto Principle. The Pareto principle states

the phenomenon that for many events, roughly 80% of the effects come from 20%

of the causes [8]. We assume that the value of information also accords with it.

 Communication between people is directed. Although the relations of people are

undirected, the information flow has a direction. Directed communication can

describe the information flow better.

 People will forward information that is valuable and consistent with their

opinion. In our daily life, people like the information of high inherent value and

consistent opinion with them. In addition, they are also willing to share it.

 Once a new technique emerged, the network based on the former technique

will begin to decay. Usually, the new technique is better than the old one. Thus

people are more willing to use the new technique. The old technique will not decay

right away in real world. But the time difference is usually small, so the assumption

makes sense.

3. Topology of Information Networks

In this section, we will study the topology of information networks and construct

multi-layer networks of information flow. First, we introduce two types of

communications in information networks. Then we define the nodes and the edges in

our models. Furthermore, we use these two types of communications to build a single-

layer network which represents the network of a certain kind of communication

technique. Finally, we obtain the multi-layer networks with the consideration of several

communication techniques being used at the same time.

3.1 Two Types of Communications

In the information networks, there are mainly two types of communications: Mass

Communication and Interpersonal Communication. Mass Communications are the

communications between media and the publics. And Interpersonal Communications

are the communications between people. [6] We will discuss the two types of

communications because they have significantly different topologies.

Team # 49436 Page 5 of 22

3.1.1 The Mass Communication

The Mass Communications developed quickly as technology improved. The first

medium come with the emergence of newspaper, which created a brand new way for

publics to gain information. With the inventions of the radio and television, media could

reach a wider range of publics. And as Internet went into more and more households,

media provided publics with varieties of information. One of the characters of Mass

Communications is that a medium can reach many people in a star topology, so this

kind of topology has a great ability of spreading information. Since media are connected,

the topology of the Mass Communication is as follow:

Figure 1. The topology of Mass Communication.

In this figure, dots represent publics and triangles represents media. Edges

represent the communications between publics and media. Nowadays, these media

include broadcasting companies such as ABC (American Broadcasting Corporation),

newspapers such as The Times and webs that have relatively great influence.

3.1.2 The Interpersonal Communication

The Interpersonal Communications always exist in our society. The very simple

example of Interpersonal Communications is speech. But as the techniques improved

quickly, the Interpersonal Communications became rich in forms from speech, mail,

telegraph, telephone to nowadays instant messages. Abundant techniques make

Interpersonal Communications much stronger and more complex. The topology of

Interpersonal Communications has random-like edges linking the person nodes. The

figure below shows a topology of Interpersonal Communications.

Figure 2. The topology of Interpersonal Communication.

3.1.3 Nodes

As we discussed above, we divide nodes into two types: one is media node, the

other is person node. Media nodes represent the media that are the centers of Mass

Communications. And person nodes represent publics. In our models, all of the nodes

have four attributes:

 The type. The type of a node shows that this node is whether a media node or

person node.

 The degree. The degree of a node is the number of nodes that connect to this

node.

 The state. A node can be in one of three states which are

剩余21页未读，继续阅读

评论收藏

内容反馈

版权申诉

pk_xz123456

粉丝: 2284
资源: 2353

社会网络中信息流与舆论演变的多层动态模型研究

SAS在多层模型中的应用

基于BP神经网络的人工单价动态预测模型研究.pdf

多层线性模型软件

HLM多层线性模型教程

多层网络中谣言传播的动态控制策略分析.docx

多层网络平均路径_双层网络_多层网络_BA_ER_

基于BERT多层网络的医疗实体抽取模型研究.pdf

基于BP神经网络的传感器网络动态采样模型研究.pdf

多层统计分析模型PPT课件

多层网络的结构与功能研究进展.docx

基于元胞自动机与多层前馈神经网络的电力信息物理融合系统安全评估模型.pdf

TensorFlow实现MLP多层感知机模型

光学多层膜热应力模型研究

河道浅滩演变预测的人工神经网络模型.pdf

多层感知器信用评价模型研究1

基于上下词及语境的CBOW多层神经网络分类模型研究源代码+模型+数据集，为短文本推荐合适的emoji

matlab开发-多层感知神经网络模型与反向传播算法

基于卷积神经网络与双向长短时记忆网络组合模型的短时交通流预测.pdf

利用Elman神经网络建立预测模型研究

基于三流卷积神经网络模型的图像分类方法.pdf

基于BP神经网络的信息系统综合评价模型的研究

基于多层感知神经网络的糖尿病并发症预测研究.pdf

神经网络与deep learning 学习与实践，多层前馈神经网络与神经网络的实现

数据回归-基于Bayesian的空间多层线性回归模型的参数估计及其应用.pdf

糖熏鸡腿颜色快速精准识别的多层卷积神经网络模型研究.pdf

论文研究-基于多层结构模型的生物信息分析平台研究.pdf

最新资源