Model of strong and weak connections of users in social networks
DOI №______
Abstract
The constructed model of strong and weak user connections in social networks, with the following assumptions:
1. The strength of the influence of one factor on another along a given path depends on the length of this path (that is, the number of edges in it).
2. The more parallel actions (in different ways) between the factors exist, the stronger the influence between them.
The impact analysis method is based on the following assumptions:
1. The strength of the influence of one factor on another along a given path depends on the length of this path (that is, the number of edges in it).
2. The more parallel actions (in different ways) between the factors exist, the stronger the influence between them.
To compare different strategies for determining user impacts, various options for the evaluation function are considered. More detailed characteristics of the interaction of factors were revealed using fuzzy cognitive maps. In problems of dynamic analysis, fuzzy values are attributed not only to relationships, but also to factors. Moreover, unlike bond weights, which are considered constant during the analysis, the quantity, factor is the value of some function, which depends on the weights of the incoming edges and the values of factors that change with time. The vector of values of all factors of the situation at a certain time forms the state of the situation. The set of edge weights is given by the adjacency matrix of the graph. The presence of a factor in a factor makes it possible not only to evaluate the power of influence on the factor, but also to express the result of the total effects in the form of a specific value. The concept of the state of the situation allows us to talk about the development of the situation in time under the influence of various external influences, expressed in a change in the values of factors, that is, to set forecast tasks (direct task), and also to explore the possibilities of managing the situation, that is, to look for influences leading to the desired (target) state (inverse problem). A model of a fuzzy cognitive graph of impacts and edge weights is considered. The solved tasks: obtaining a forecast of the situation (direct task) and finding control actions (inverse problem).
Keywords: model; method; analysis; graph; strategy; characteristics; influence; fuzzy; interaction; computation; path; vertex; edge; function; result; consonance; variable; assumption; monotonous; estimated; total; dynamic; cognitive maps; set; gain; task; module, factor; transitive.
References
1. Kosko B. Fuzzy cognitive maps // International Journal of Man-Machine Studies, 1986. Vol. 1. P. 65–75.
2. Maruyama M. The Second Cybernetics: Deviation-Amplifing Mutual Causal Processes // Amer. Scientist. 1963. Vol. 51. P. 164–179.
3. URL: http://posp.raai.org/data/posp2005/Kuznetsov/kuznetsov.html.
4. Liu Z.-Q., Zhang J. Y. Interrogating the structure of fuzzy cognitive maps // Soft Computing, 2003.
5. Кофман А. Введение в теорию нечетких множеств. Москва: Радио и связь, 1982.
6. Поспелов Д. А. Нечеткие множества в моделях управления и искусственного интеллекта. Москва: Наука, 1986.
7. Pappis C. P., Adamopulos G. I. A computer algorithm for the solution of the inverse problem of fuzzy systems // Fuzzy Sets and Systems, 1991. V. 39. P. 279–290.
8. Bourke M. M., Fisher D. G. Solution algorithms for fuzzy relation equations with max-product composition. // Fuzzy Sets and Systems 1998. V. 94. P. 61–69.
9. Марковский А. В. О решении нечетких уравнений типа «max-product» в обратных задачах управления и принятия решений // Автоматика и телемеханика, 2004. №9. С. 149–159.
10. Markovskii A. V. On the relation between equations with max-product composition and the covering problem // Fuzzy Sets and Systems, V. 153. Issue 2, 16 July 2005, P. 261–273.
