Studies on deep learning techniques For the analysis of social networks

Abstract

Recently, social networks are used for many activities such as blogging, trading, sharing information and so on. One of the most significant issues in social media network is Influence Maximization (IM), where the influential nodes are needed to be determined for several applications namely network monitoring, product/brand recommendation and so on. This problem is specifically handled through the designing probabilistic model where the influence spreads over the network is determined for influential nodes tracing. But, the sensitivity and specificity remains a major concern while increasing influence maximization. Therefore, the proposed research work focused for increasing the sensitivity, specificity and accuracy involved for influential node tracing in social network. newlineA Tuned Linear Threshold Model (TLTM) is proposed for tracing the influential users or nodes in social network with better accuracy and lower response time. TLTM is proposed by using Node Extraction Algorithm, Feature Selection Node Algorithm and Tuned Linear Threshold Algorithm. At first, data are retrieved from the database and then the Node Extraction Algorithm is applied to extract the nodes based on the location and significance of each node. Followed by this, Feature Selection Node Algorithm is employed to extract the relevant or best nodes by computing the fitness value depended on the lower bound and upper bound for obtaining influential nodes. Lastly, Tuned Linear Threshold Algorithm is used to estimate the influence spread for tracing the influential nodes in the network with better sensitivity. newline