Clustering analysis in social network using Covering Based Rough Set

This paper focuses on solving the problem of classification and clustering in social network by using Rough Set. When the data set consists of missing or uncertain data then the Rough set is proved to be an efficient tool. To solve a problem under the domain of social network, the problem must satisfy the fundamental property of rough set i.e., the attribute of the problem must holds true for equivalence relation. Hence, before implementing rough set to the specific problem of social network, it must be redefined in a way that properties of transitive, symmetric and reflexive should holds true. In this paper, we have studied on the concept of Fiksel´s societal network and used it for redefining the social network problem in terms of equivalence relationships. Further, we had defined the Social network in terms of graph theory and mathematical relations. We had proceeded further in defining the Fiksel´s societal network and social network with respect to rough set. Fiksel had defined the social network in terms of structural equivalence. We have discussed on the limitation of Rough set and observed that use of Covering Based Rough Set as an extension of Pawlak´s rough set seems to be a better alternative. There are six types of covering based rough set. To keep continuity in this paper, we have mentioned about Covering based rough sets. Covering based rough set extends from partitioning in rough sets to covering of the universe and is flexible, when compared with rigid equivalence relation.