The paper presents a description of classical and modern multimedia information resources security threats in social networks. It also provides an overview of methods and models for analyzing social networks. Proposed an approach to the creation of a methodology for building security systems for the exchange of multimedia content in social networks through the development of conceptual foundations, methods, and technologies for detecting, assessing, and countering information security threats.
The continuous development of global communication services has turned social networks into the dominant segment of information exchange and citizen communication in the virtual space.
The availability of personal media for generating multimedia content and online communities for
sharing it has led to explosive growth in the volume of multimedia information that circulates in
social networks. In contrast to textual information, for the processing of which there are a wide
number of tools and methods [
The growth in the number of users of social networks and multimedia services is currently so active that it outstrips the increase in the world’s population. In Fig. 1, it can be seen that the growth rates of the number of users of Facebook and YouTube are significantly higher than the rate of population growth (151.87 and 111.64 compared to 83.873). The increase in information security threats in social networks is directly related to the emergence of such an effect. This can be explained both by an increase in the number and complexity of the content structure and by a significant increase in the number of inexperienced (unprepared, new) users.
Information security threats are of a different nature and can be implemented using special
software or technologies that are inaccessible to the vast majority of users of social networks due to
their lack of appropriate professional competence [
Information security threats in social networks l a i c i f i t r a g n in io t a n e ip r c o - ic g l
b in u fr p e h t o r t s A h g u o r th n n io itlaou tram
o ip f n in a M l a eh re i c t d so t l
u o e o h tinh ityhw itrhg ilrsa itn opy e u c t m omm tfeh a g c o in k tn ts r o o e P tw sn en co , f n o litrraecpoduo tiitirrsdbuno ltlireaeabm n o a titIenon tiliaconp tijcenobo u d
Informational and psychological impact on group users by imitating the mass discussion of any information (the so-called astroturfing). This type of threat has a negative impact on the joint work of the group members, reflecting on the consciousness of the participants in the discussion.
Manipulation with information (information overload, misinformation, distortion of information, or mixing true facts with false ones). This threat is capable of affecting the mind and consciousness of the group members, as well as inciting them to carry out destructive and socially dangerous actions.
Posting materials in the social network community without the consent of the copyright holder. The ease of infringement of copyright in community-posted student work leads to unauthorized g n i y l l u b r e b y C copying and use of fragments of the work.
Reproduction, duplication, or distribution of objectionable material intentionally posted by group members.
Cyber humiliation and cyberbullying of network members. Humiliation and insult in community discussions destroy the moral and psychological atmosphere and hinders the implementation of activities.
In [
To effectively reduce the degree of influence of threats to information security in social networks, it is necessary to adequately apply methods of analysis of social networks, which make it possible to identify potential “centers of influence” capable of forming a collective public opinion for destructive purposes.
Currently, in the analysis of social networks, there are four main areas of research [
link analysis
Structural and behavior
analysis of social networks
Definition of communities on social network
In the structural approach, all network participants are considered as nodes of the graph, which affects the configuration of edges and other network participants. The main attention is paid to the geometric shape of the network and the intensity of interactions (the weight of the edges); therefore, such characteristics as the mutual arrangement of vertices, centrality, and transitivity of interactions are investigated. Structural and network exchange theories are used to interpret the results in this direction.
The resource-based approach considers the possibilities of participants to attract individual and network resources to achieve certain goals and differentiates participants who are in identical structural positions of the social network by their resources. Knowledge, prestige, wealth, race, gender can act as individual resources. Network resources are understood as an influence, status, information, capital.
The normative approach examines the level of trust between participants, as well as the norms, rules, and sanctions that affect the behavior of participants in a social network and the processes of their interactions. In this case, the social roles that are associated with this edge of the network are analyzed, for example, the relationship between the manager and the subordinate, friendship, or family ties. The combination of individual and network resources of a participant with the norms and rules in force in this social network forms his “network capital.” In a simplified form, “network capital” can be viewed as the sum of some advantages that a participant can receive at an arbitrary point in time to achieve a certain goal.
A dynamic approach is a direction in the study of social networks, in which the objects of research are changing in the network structure over time: for what reasons do the edges of the network disappear and appear, how does the network change its structure under external influences, are there any stationary configurations of the social network.
As can be seen from previous, when analyzing social networks, a fairly wide range of problems is solved and methods from various fields of knowledge are applied. However, to directly identify the characteristics and properties of social networks and their segments in order to identify potential sources of threats to information security, it is necessary to consider specific models for analyzing social networks (Table 1).
Content and characteristics of the model For many social tasks, such as finding a job, weak ties are much more effective than strong ties (ant colony optimization approach) The hypothesis is that each person is familiar with any other inhabitant of the planet through a chain of mutual acquaintances, on average, consisting of six people. So far, this claim has not been refuted. On the contrary, as proof of the correctness of the hypothesis, the observation is put forward that the diameter of most networks is relatively small.
Any social network can be mathematically represented as a graph (stochastic block models, probabilistic graph models, conventional graph models) To determine the relative importance (weight) of the graph vertices (that is, how influential a participant is within a particular network), the concept of centrality is introduced—a measure of proximity to the center of the graph.
Designed to separate informative web pages from spam. TrustRank is a value that gives an estimate of whether a particular site can be trusted, assuming that it does not contain spam. The more links there are on the site, the less trust is “passed” through each such link. TrustRank decreases with increasing distance between it and the original sample.
The strength of the structural position is the main indicator that determines the differences in the resources of network participants.
Community analysis allows studying the stability of social structures.
The simplest case of a linked group is a community where each member is associated with everyone, and other members of the network cannot be included in this group, since they do not have Data dimensionality reduction methods Structural equivalence of network participants Role algebras Analysis of dyads and triads Stochastic models Network dynamics models Algorithmic tools for analyzing the evolution of networks Analysis of network development graphs Predicting bond formation Ontology-based models connections with all members of the community.
The projection of the network vertices into the Euclidean space of reduced dimension is considered to describe the relationships between the rows and columns of this matrix. As a result, it is possible to visualize changes in the status of a network user against the background of changes in the statuses of subgroups.
The approach is the opposite of exploring related groups.
Participants are equivalent when they occupy the same positions in the social structure of the network, that is, when the structure and type of interaction of these participants with others are equivalent, while the equivalent participants in the network should not interact with each other.
The direction of analysis of social networks, which focuses on identifying the logic of interactions of network participants in block models, makes it possible to identify similarities in the principles of relationships between participants in various social networks Dyads are a set of two network members (vertices) and all interactions (edges) between them. The dyad for each type of interaction can be in one of four states: there is no connection between the participants, the connection is directed from the first participant to the second, the connection is directed from the second participant to the first, mutual connections of the participants. The analysis of dyads helps to establish the probability of the presence of an edge between them, the degree of dependence on the properties of participants, to determine the conditions and directions of information transfer, etc. For triads (three interacting participants), questions of transitivity of interactions are additionally investigated.
The main idea behind probabilistic directed graph models is that each social network can be viewed as a realization of a random twodimensional binary array.
Consider the strength and speed of changes in connections in social networks (graph evolution models, “closed triangle” model, “forest fires” model) The algorithms are based on dynamic programming, exhaustive search, maximum matching, and greedy heuristics. The main focus is on identifying approximate user clusters and their temporary changes.
Approaches to the analysis of network evolution based on the paradigm of association rule mining and frequent pattern mining.
The rules of the evolution of the graph, a new type of frequency models are introduced, and the problem of finding typical models of structural changes in dynamic networks is considered.
Graph evolution models are usually created to evaluate the general statistical properties of existing graphs. You can also try to calculate whether two specific vertices will be connected to each other after a certain period of time. This is a computational task based on the analysis of the evolution of a social network in time and is called the problem of predicting connections.
Parameters of social networks can be evaluated using ontologies.
First, an analysis of the types of network elements is performed:
people, objects (music, photos, videos, messages), interactions
(knows, reports, comments, etc.). Then the authors use the existing
ontology resources and additional options for all kinds of
relationships. The FOAF ontology is used to identify the members of
a social network and the content they add to the network. A new
version of SCOT is used to describe tags.
* Summarized from [
Among the many existing models for analyzing social networks, the most interesting are those that can actually be implemented on the basis of available statistical information (metrics of a social network and its participants).
One of these characteristics is the “level of trust.” The algorithm for calculating the level of trust
(TrustRank) [
To strengthen the results of calculating the level of trust, it is advisable to use the indicator
“Strength of the structural position of a participant,” which determines the differences in the resources
of network participants. In the theory of network exchange, to measure this characteristic, [
It is advisable to track the spread of malicious information from centers of influence using
“Network Dynamics Models” [
Graph evolution models [
Triangle-Closing Model [
The model of “forest fires” [
In addition, studies [
However, to build an effective security system for multimedia content in social networks, it is necessary to develop an algorithm based on the use of available information about participants, groups, and communities. This will allow developing a mechanism for identifying threats and identifying the sources of their occurrence.
To develop a methodology for constructing security systems for multimedia information resources in social networks, the concept of modeling the behavior of members of a social network is proposed, which is implemented at three levels.
At the first level, an element of a social network (agent or actor) is determined, the behavior of which forms the behavior of the social network as a whole, and is decisive for the study and construction of a security system. At this level:
1. A set of actions of an individual element of a social network is determined, which together form behavior.
At the second level, models of collective behavior (group dynamics) are built. At this level:
2. The dynamic characteristics of the group’s behavior are determined—stability, coordination of actions, self-organization of the group, temporal characteristics of the dynamics of behavior.
The formed complex of tasks, separated by levels, makes it possible to form an economically grounded methodology for constructing a security system for multimedia information resources of social networks.
The methodology construction process consists of five stages:
5. Development of methods for determining the most likely threats and assessment of their cost indicators.
The inclusion of economic indicators in the model for analyzing the security of multimedia information resources in social networks will expand the capabilities of detecting threats and sectors of influence (critical nodes) in social networks.
The use of standard models of analysis of social networks at the present stage does not allow identifying the sources of threats to the security of multimedia information resources in social networks. This is primarily due to the fact that the main tool of influence is multimedia content, which, unlike text content, is not subject to indexing to increase search speed. Mass identification of multimedia content in modern conditions is possible only on the basis of the name of the file itself and hashtags on the corresponding page. This problem critically affects the development of effective models for analyzing multimedia information resources in social networks and requires the development of a new methodology for identifying and countering information security threats. 6. References