Dynamics of online social network based on parametric variation of relationship

The current research in opinion mining is largely based on content analysis of social interactions of users on a network. However social interactions are also governed by relationships existing between the various nodes. The role of relationship specific attributes on categorical influence prediction in a social network forms the basis of the presented work. This paper proposes a two phased collaborative model for predicting spread of influence in a social network by utilizing multiple relationship specific parameters. The initial phase identifies and visualizes the varied opinions based on relationships in a network which are then quantified through a distinct measure, the opinion metric, in the second phase. The metric takes in consideration the opinion and page rank centrality of respective nodes to generate the strength of node's negative influence factor. A high value is indicative of higher probability of spreading maximum negative influence. An experimental study on a sample network of ten nodes is conducted using Gephi, a social network analysis software. Using the opinion metric, the node capable of spreading the maximum negative influence in the network was identified. Early identification of malicious nodes in a network can be of immense help in various sectors such as Marketing, Defense, Stock markets, IT industry etc.