Abstract:
Social networks usually have underlying communities and it is the task of community detection
algorithms to detect these communities. Social networks consist of various nodes each playing
a different role in creating communities.The property we will focus on in this thesis is influence.
Influence has many definitions in different scenarios. In some cases the nodes which display
the most centrality are considered the most influential nodes where as in other cases nodes the
betweeness of a node defines its influence. In social networking websites the influence of a node is
measured by how many nodes it is connected to. It can commonly be observed that celebrities with
a high number of followers have the most impact on facebook. This influence comes in use when a
product needs marketing or when some cause needs to be promoted. Hence we focus this study to
firstly find out the impact of the most influential nodes in a social network and then find out the
communities they create.
Our algorithm starts with the top-1 node and moves down the nodes in a descending order based on
the degrees of the nodes and classifies them into communities.
