A Novel Approach for Secure Hidden Community Mining in Social Networks using Data Mining Techniques

Abstract

network community contains a group of nodes connected on the basis of certain relationships or same properties. Sometimes it refers to the special kind of network arrangement where the Community Mining discovers all communities hidden in distributed networks based on their important similarities. Different methods and algorithms have been employed to carry out the task of community mining. Conversely, in the real world, many applications entail distributed and dynamically evolving networks. This leads a problem of finding all communities from a given network. Detecting evolutionary communities in these networks can help the user for better understanding the structural evolution of the networks. In this research, first a new bipartisan scheme using k- Dimensional (KD) -Tree to deal with the recursive bisection method is proposed; next an Improved KD-Tree algorithm to deal with the multidimensional problem is put forward. The security issue such as a Sybil attack (Multiple fake Identities attack) arises in these network structures. It can be mitigated by fixing the target time by using SICTF (Sybil Identification using Connectivity Threshold and Frequency of visit) algorithm. The problem faced by the mining community of heterogeneous network can be addressed by using Convergence aware Dirichlet Process Mixture Model (CADPM). KeywordsMining, Links, Nodes, Social Networks, Sybil,