Abstract
The ability of social networks to disseminate information across
individuals and groups is based on the social influence that people have
on each other.
In this context, the so-called influence maximization problem consists
in identifying the most influential nodes of a given social network.
This problem has practical relevance in a wide range of applications and
despite the underlying computational complexity, several solution
techniques have been presented in the related literature. Nonetheless,
bringing together technical feasibility and satisfactory accuracy is
still a challenging open research issue.
In this work, we address the influence maximization problem with two
complementary contributions. First, we analyze two well-known influence
diffusion models, namely the independent cascade and the linear
threshold models, and provide a new methodology for addressing the
problem of computing the influence spread in a given network.
Subsequently, we propose a novel approach to select an initial set of
nodes that optimizes the influence spread in large-scale scenarios.
We apply these techniques to several large-scale experiments in
real-world scenarios achieving results comparable with the
best-performing state-of-the-art algorithms in a shorter computational
time.
This work has been submitted to
the IEEE for possible publication. Copyright may be transferred without
notice, after which this version may no longer be
accessible.