Copula Tubular Quantum Based Spearman Deep Neural for Influential Node Tracing in Social Network

Abstract

Social networks are one the foremost origins of information transmission at present. Nevertheless, not all nodes in social networks are indistinguishable. As a matter of fact, certain nodes are said to be more influential than others, or to be more specific, their information gravitates to proliferate more. Identifying the most influential nodes in a social network called as the Influence Maximization problem remains one of the hot issue with the evolution of Internet and social media. Several methods have been proposed to identify influential nodes in composite networks, ranging from parallel algorithm to distribution difference and non-overlapping communities. However, most of the previous methods do not take into account error involved in overlapping communities in social network. To address this issue, in this work, a Copula Tubular Neighborhood and Quantum-based Spearman Deep Neural Network (CTNQ-SDNN) for influential node tracing in social network is proposed. The CTNQ-SDNN method is split into three sections. First, computationally efficient nodes are identified by employing Copula Probability Node Pre-processing model. Then, with the identification of effective nodes, accurate feature extraction is made by means of Tubular Neighborhood Intensity-based Feature extraction model. Finally, influential node selection and tracing in case of tie or overlapping communities with the purpose of reducing error rate is proposed by utilizing Spearman Deep Neural Influential Node Tracing model. Experimental results on telecom dataset social networks, comparing our proposed method against state-of-the-art methods in current literature, indicates our method to be efficient and robust in tackling the influence maximization issue even in case of tie or overlapping communications in social network via error rate.