Maximizing Spread of Influence in Social Networks

Please enable javascript in your browser.

Doroud, Mina | 2009

531 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 39579 (02)
University: Sharif University of Technology
Department: Mathematical Sciences
Advisor(s): Mahmoudian, Ebadollah; Rabiei, Hamid Reza
Abstract:
Social networks play a fundamental role as a media for the spread of inuence among its members. How people like to adopt innovation from their friends is known as word-of-mouth eect and has a long history of study in social science and recently in computer science and mathematics. In this context, The Inuence Maximization Problem is about nding initial active set of specic size in order to maximize the nal adoptions in the network through diusion of inuence. The optimization problem of the most inuential nodes selection is NPhard. As a result, some heuristic algorithms are needed to approximate the nal result. In this thesis some new heuristic algorithms were proposed with respect to centrality concept in the graph theory. They are compared with themselves, also with the traditional maximization algorithms such as simple greedy or degree centrality, along with some other heuristics like set covering algorithm. For testing the result, dierent social networks were used, like weblogs, world wide web and Iranian mathematician collaboration social networks. The latest one where created and observed to establish a real social network with more than four thousands nodes.
Keywords:
Social Networks ; Virul Marketing ; Social Influence ; Social Influence Spread