Link Performance in Community Detection Using Social Network

The structures of social network provide the regularities in the patterning of relationship among social entities. The relationship can be determined by frequency of interactions, propagation, and cultural activities and strong/week ties in social network. The structure of social network characterizes in terms of nodes (individual actors, people, or a thing within the network), ties, edges and links. The node is to be optimized to improve the link availability, clique through rate, effectiveness of link usage. The node optimization could be achieved by optimization of links. There are many predicting techniques are used to predict the nodes decision to improve the ‘influence propagation’ and ‘retrieve information’ from the requested users. The existing link prediction algorithms are based by the performance of the community’s with multiple parameters in the social network environment, and each parameter has to be quantified for best trade among the links. Hence, a node proximity clustering (NPC) algorithm has considered multi-dimensional parameters along with node weight value to quantify the importance value of the link prediction. The performance of the designed algorithm has been experimented in the various scenarios. The node proximity clustering algorithm (NPC) has been experimented in the defined environment with the fast greedy (FG), binomial (B), and socio rank (SR) of prediction algorithms. The results had been compared and analyzed. The proved algorithm has yielded 97.66% accuracy. It has validated with improvised label propagation algorithm and found that the proved algorithm accelerated the accuracy up to 15%.