DonNTU   Masters' portal

Irina Shcherbina

Faculty of computer science and technology (CST)

Department of Department of computer modeling and design (CMD)

Speciality «Information technologies in media industry and design»

Research of functional features of social networks for the development of tree-like models of Internet user interaction

Scientific adviser: Ph.D., V.Pavlii

Resume Abstract

Abstract

Attention! At the time of writing this abstract, the master's work is not completed. Estimated completion date – May 2021. The full text of the work and materials on the research can be obtained from the author or his scientific adviser after the specified date.

Содержание

• Introduction
• 1. Theme urgency
• 2. Goal and tasks of the research, expected outcomes
• 3. Overview of research and development
• 3.1 Overview of international sources
• 3.2 Overview of national sources
• 3.3 Overview of local sources
• 4. Social media analysis
• 4.1 A brief history of the development of social networks.
• 4.2 Areas of analysis of social networks.
• 4.3 Model. Opinion, influence, opposition of opinions. Social network structure.
• Conclusion
• References

Introduction

Social networks are one of the brightest phenomena of modern information and communication technologies and a generally recognized trend in their development. Here are just a few examples.[1]:

• every second 8 people on the planet become part of any of the existing social networks;

• half of all users spend one to five hours a week chatting on social networks;

• Facebook is the 3rd largest country in the world, after China and India, with a population of about a billion people;

• the chances that the average person under thirty is on a social network is over 50%.

Social networks are firmly entrenched in people's lives. It's hard today imagine a person who is not registered on social networks. But social networks are not only those sites that we regularly visit in internet. Any group of people interacting with each other forms social network. In science, this trend has led to the emergence of a new directions for research.

1. Theme urgency

Social networks are social structures consisting of a set of agents (subjects – individual or collective, for example: individuals, families, groups, organizations) and a set of relations defined on it (a set of connections between agents, for example: acquaintance, friendship, cooperation, influence, communication ) – have been the subject of active research since the second half of the 20th century. When modeling social networks, it becomes necessary to analyze them, including as networks of influence – taking into account the mutual influence of network members, the dynamics of their opinions.

With the development of information and telecommunication technologies over the past ten years, the importance of a new type of resources – online social networks – has grown significantly as a means of disseminating opinions that affect the actions of network users. Social media researchers (such as Jackson M and Roberts F.) do not consider management problems. The results known in the theory of management of socio–economic systems (Kononov D.A., Kononenko A.F., Kulba V.V., Novikov D.A., Chkhartishvili A.G., Makarov V.L., etc.) information management mechanisms (influencing the awareness of system participants) do not fully take into account the specifics of social networks [2], which determines the relevance of the topic of the master's work.

2. Goal and tasks of the research, expected outcomes

The purpose of the work is to study the functional features of social networks and to develop models of user interaction.

Main research objectives:

1. Revealing the specifics of social networks as objects of management; formulation and classification of information management tasks in social networks.
2. Development and research of models and methods (mechanisms) of user interaction in social networks, including:

   – models of information influence in social networks, including taking into account the reputation of participants;

   – models and methods of information management in social networks;

   – models and methods of information warfare in social networks;

   – models of information epidemics in social networks and methods of protection against them.

3. Creation of a software package for research using the developed models and methods.

The main research method is mathematical modeling using the approaches and results of game theory, the theory of active systems, decision theory and operations research.

3. Overview of research and development

Currently, the analysis of social networks is one of the most intensively developing areas not only in sociology, but also in other technical disciplines. The interest in them is dictated by the fact that the fact of the special position of this object of research is becoming quite obvious, entailing a new set of explanatory models and analytical tools that are outside the framework of conventional research methods – both quantitative and qualitative.

3.1 Overview of international sources

Over the past 10 years, research has been actively carried out in the study of social networks, which give rise to new and effective models and methods for their assessment. I would like to highlight some foreign experts on this subject: Harary F. [ 3 ], French JR, Harary F. [ 4 ], De Groot MH, Friedkin NE [ 5 ], Roberts F. and others. To date, many works are devoted to the study models of social networks, including models of influence. In some of them, more attention is paid to the network of influences itself, in others – to the processes and rules of interaction.

3.2 Overview of national sources

The issue of researching social networks did not go unnoticed among Russian scientists. These include Shvetsov D.A., Kononov D.A., Muromtsev V.V., Ponomarev N.O., in their research work they consider the technologies and methods of the Internet social network as a means of modern communication [ 6 ]. You can also highlight the scientific research of Gubanov D.A., Novikov D.A., Chkhartishvili A.G., devoted to the model of information influence, control and confrontation [ 7 ].

3.3 Overview of local sources

As a result of a review of the work of students of Donetsk National Technical University, it was revealed that some students also dealt with issues related to social networks:

1. Kucherov P. Analysis and development of the social network for distance education and preparation to Institution of higher learning of schoolchildren [8];
2. Plotnikov D. Methods and means of improving the efficiency of Internet—based applications [9];
3. Vestnikova A. Theme of master's work: Research and analysis of models of information retrieval, their use on the site Online services in social networks. [10];

However, the issues of researching social networks have not been covered, although they have a number of significant differences associated with the methods and ways of studying them.

4. Social media analysis

4.1 A brief history of the development of social networks

Relations between people, the resulting structures of relations, their analysis and management, related issues of power worried people from the first steps of organizing society. With development society, these issues began to be studied on a systematic basis. As an example, we can cite the well-known School of Pythagoras: “Everything began with the stages of teaching moral purity. He explained the laws of man's relationship – with man, with nature, with God " [11]. However, the actual term "social network" was first introduced by John A. Barnes only in 1954 in [12].

Computer social networks have arisen as a result of advances in information and communication technologies, among which the following can be noted:

• 1971 – Email (a computer-based social network), Ray Samuel Tomlinson;

• 1988 – "IRC" (Internet Relay Chat) – communication in real time, Finnish student Jarkko Oikarinen;

• 1991: Internet, Timothy John Berners-Lee (Sir Timothy John “Tim” Berners-Lee).

ПAll of the above formed the infrastructural basis of modern social networks. The actual networks in their modern understanding did not arise so long ago, and the landmark events here were the creation of the Classmates.com network by Randy Conrad in 1995 and the creation by Mark Elliot Zuckerberg on Facebook in 2004.

The real world and the virtual world do not exist separately from each other. An example is the Groupon network, created in 2008 – connecting communication and business, on-line and off-line worlds. Many researchers and leaders of the high-tech industry (for example, Google CTO Ray Kurzweil) see the future of social networks as mixed networks, where people and robots communicate, develop, and solve problems together.

In the analysis of social networks, two approaches can be distinguished, which can be conditionally called look outside and look inside. The first approach is more common nowadays. This is partly due to the fact that from a mathematical point of view, a social network is a sufficient a well-studied object – a graph that is a collection of nodes or vertices and links or edges between them. This view allows you to visualize, make clear, the state of the network or its part and understand some of the processes taking place in it.

4.2 Areas of analysis of social networks

According to J. Scott [13], there are many different directions in the development of modern social network analysis, appearing at different times and in their totality forming a kind of "pedigree" of this method. On this "family tree" there are four main lines: 1) Gestalt theory, 2) field theory and sociometry, 3) group dynamics, 4) graph theory. These lines of modern social network analysis have served as the foundation of sociometric analytics, which has spawned many technical achievements using methods of graph theory.

Structural and functional anthropology, research and theoretical searches represent an independent line on the "family tree" in which they stimulated the formation of both sociometry and network analysis (Warner, Mayo, Gluckman). They were born at the intersection of two lines in the middle of the 20th century. theoretical and methodological achievements of Homans, Barnes, Bott and Nadelya. Group dynamics played an important role as the primary source. She also influenced the emergence of modern analysis of social networks. It is necessary to add the names of Radcliffe-Brown, Haider, Koehler, White, Moreno, Newcomb, etc. to this scheme (Fig. 1)

The term “social network” was introduced into scientific circulation in 1954 by sociologist D. Barnes in his work “Classes and Meetings in the Norwegian Island Parish”, published in the collection “Human Relations”. D. Barnes developed J. Moreno's approach to the study of relationships between people using sociograms. It is no coincidence that anthropologists became pioneers in the development of the methodology of modern network analysis, for whom it was important to fix the differences in the structures of small and traditional societies and to develop methods for their empirical study.

Radcliffe-Brown was the first to use the terminology of network research and urged to consider the social structure as a network of social relations. In the UK to study social structure Radcliffe-Brown used two methods – morphological and physiological study of social systems. The functions of the first include the definition, comparison and classification of various structures. Physiological tasks method – the study of the mechanisms that support the existence of a system of social ties. “Social physiology ... – stresses Radcliffe-Brown, – deals not only with social structures, but with all types of social phenomena. Morality, law, etiquette, religion, governance and education are all part of a complex mechanism through which the social structure exists and remains intact. If we take a structuralist point of view, we will see that we are studying all these things are not abstract and not isolated, but in direct and indirect interaction with the social structure, i.e. we constantly consider how they depend on social relationships between individuals and groups, as well as how they influence this relationship. ”[14]

By the beginning of the 1990s, a dual situation began to develop in network analysis: a substantial amount of empirical works had been accumulated, methodology was actively developed and methods were improved. But at the same time, there were no theoretical works in which the would be a network analysis concept.

At the present stage, most of the research carried out is devoted to the problems of forming communities on the Internet, building them structures, identifying clusters. In addition, studies of specific subject communities are widespread.

4.3 Model. Opinion, influence, opposition of opinions. Social network structure.

A social computer network means a structure that consists of many agents (users) and a set of relations (influences between agents) defined on it. Formally, a social network can be represented as a graph G (N, E), in which N = {1, ..., n} is the set of vertices (agents), and E is the set of edges (influences), which reflect the interaction of agents.

Influence is understood as the process of influence by one agent (by the user) on the opinion of another agent and, as a result, a change opinions of the latter. To date, many models have been compiled influence on social networks.

By interacting in computer social networks, some agents influence the opinions of others, thereby prompting the latter to certain actions. In the present the work has built a model that is as close to reality as possible, since recently this problem has become relevant. This model will serve as a basis for further research.

The main models of social networks can be divided into the following groups:

1.1. Models with thresholds. The main feature of this class of models is the presence of thresholds Фj [0,1], which can be both linear, and nonlinear. The agent in this model is represented as a social node networks. It can initially be in only one of two states: active or inactive. Moreover, it is usually assumed that the transition is only from inactive to active. Agents interact and influence each other friend with certain degrees of influence Wij. If active agent i affects on an inactive agent j with the magnitude of the influence Wij, then agent j is activated when the following activation condition: Wij> Фj.

The threshold value depends on the mathematical model itself: in one the threshold values can be fixed for all agents, in another determined randomly according to some probability law distribution, in general, individual differences are based on personal qualities of the agent, his interest, etc.

1.2. Models of independent cascades. In this model group, network agents also can be in only one of two states. Upon activation agent i at time t, with some probability Pij at the next step he gets the opportunity to activate each of his neighbors j. On in the next step, already activated neighbors with some probability can activate your neighbors, etc.

1.3. Percolation and infestation patterns. This group of models is one of the most popular ways to study the dissemination of information in social network.

The classical model is based on the disease cycle of the carrier (agent). The agent is initially susceptible to disease. When coming into contact with with an infectious agent, the agent in question can be infected with some probability. After a while, the agent recovers, thus acquiring immunity to the virus or dying. Immunity with decreases over time, and the agent becomes receptive again.

One of the main indicators of percolation and infestation patterns is the critical probability of infection of a neighboring agent – "Epidemic threshold" B, if the threshold is exceeded, then infection will spread to the entire network. The spread of infection depends on the model social network. The epidemic threshold may be absent, for example, as in networks without scale. If an infection appears in them, then it will spread to the entire network [15].

1.4. Ising models. Ising model is a mathematical model taking into account only the interaction of the nearest atoms. With her help social networks can also be modeled. A similar task was covered in article [16]. Its authors suggest that due to the model Ising can model independence in a large group. The decisive factor is the interaction between the nearest neighbors in networks. The influence of authority is an external field for a social group.

1.5 Models based on cellular automata. This group of models is also one of the popular ways to represent propagation information on the network. The social network is presented as a complex system, which consists of many agents interacting with each other. it interaction affects the collective behavior of agents, analysis or the prediction of which is very difficult.

The set of agents of the cellular automaton makes up a regular lattice. The current state of the agent is characterized by a variable, which is determined at every moment of time. The object states change according to certain probabilistic rules that depend on the nearest neighbors and, perhaps even from a chosen agent, at finite intervals.

An example of using models is shown in article [17]. In this job presents the effect of "word of mouth" in the dissemination of information in social network. The agent connects with strong bonds with the agents, members of his own network. In addition, the agent also has weak ties. With the help of them, he is connected with agents of other own networks. Although the likelihood that information will spread through weak bonds are lower than strong bonds, the authors come to the following conclusion: weak links affect the speed of information dissemination; at least as much as the strong.

1.6. Markov chain models. This model considers interactions between agents. And not only actions are modeled each agent, but also the actions of the group as a whole. Such a structure called two-tier.

Model groups 1.3-1.6 deal with interaction rules agents. At the same time, little attention is paid to the network of influence itself, its properties, structures and interaction processes, which is one of disadvantages of these models.

Modeling influence on social networks has also touched on game theory. There are a number of game-theoretic models of networks:

2.1 Mutual awareness models

2.2 Models of concerted collective action

2.3 Communication models

2.4 Network stability models

2.5 Information Influence and Management Models

2.6 Models of information confrontation

These models emphasize the relationship between agents and their awareness. The agent acts in such a way as to maximize his gain, but its benefit also depends on the actions of other agents.

References

1. Interesting facts about social networks [Electronic resource]. - Access mode: https://cso-krokus.com. ua / interesnaya-statistika / 1143-interesnye-fakty-o-soczialnyx-setyax.html
2. Gubanov D.A. Models of information management in social networks / D.A. Gubanov. - M .: Nauka, 2009 .-- 3 p.
3. Matthew O. Jackson Social and Economic Networks [Electronic resource]. - Access mode: https://web.stanford.edu/~jacksonm/netbook.pdf
4. Harary F. Graph theory [Electronic resource]. - Access mode: https://cs.bme.hu/fcs/graphtheory.pdf
5. Friedkin N.E., Eugene, C .: Social Influence Networks and Opinion Change- Advances in Group Processes, vol. 16, pp. 1–29 (1999)
6. Shvetsov D.A., Kononov D.A., Muromtsev V.V., Ponomarev N.O. Internet-social network as a means of modern communication: technologies and research methods [Electronic resource]. - Access mode: https://economics.rsuh.ru/jour/article/view/13/14 #
7. Gubanov D.A., Novikov D.A., Chkhartishvili A.G. Social networks: models of information influence, control and confrontation [Electronic resource]. - Access mode: https: // socioline. ru / book / gubanov-da-novikov-da-chhartishvili-ag-sotsialnye-seti-modeli-informatsionnogo-vliyaniya-upravl
8. Kucherov P. Analysis and development of the social network for distance education and preparation to Institution of higher learning of schoolchildren [Электронный ресурс]. – Режим доступа: http://masters.donntu.ru/2012/fknt/kucherov/
9. Plotnikov D. Methods and means of improving the efficiency of Internet—based applications [Электронный ресурс]. – Режим доступа: http://masters.donntu.ru/2012/fknt/plotnikov/
10. Vestnikova A. Theme of master's work: Research and analysis of models of information retrieval, their use on the site Online services in social networks. [Электронный ресурс]. – Режим доступа: http://masters.donntu.ru/2015/fknt/vestnikova/
11. Dialogues with Pythagoras - [Electronic resource]. - Access mode: http://en.pythagoras.name/dialogues-with-pythagoras.html
12. John Arundel Barnes. Class and Committees in a Norwegian Island Paris. - [Electronic resource]. - Access mode: http://pierremerckle.fr/wp-content/uploads/2012/03/ Barnes.pdf
13. The lineage of social network analysis Source: Scott 2000 - [Electronic resource]. - Access mode: https://www.researchgate.net/ figure / The-lineage-of-social-network-analysis-Source-Scott-2000_fig1_274265119
14. Radcliffe-Brown A.R. - Method in social anthropology (2001) - [Electronic resource]. - Access mode: https://platona.net/load/knigi_po_filosofii/antropologija/rehdkliff_braun_a_r_metod_v_socialnoj_antropologii_2001/ 5-1-0-2913
15. Romualdo P., Alessandro V. Epidemic Spreading in Scale-Free Networks // Physical Review Letters. 2001. No 14 (86). P. 3200-3203.
16. Tarnowe. Like Water and Vapor - Conformity and Independence in the Large - Access Mode: http://cogprints.org/4274/1/LargeGroupOrderTarnow.pdf
17. Goldenberg J.,Libai B., Muller E. Talk of the Network: A Complex Systems Look at the Underlying Process of Word-of-Mouth // Marketing Letters , 2001 pp. 11-34.