Machine Learning based Analysis of Influence Propagation on Social Network with Time Series Analysis

The influence propagation between the nodes is the key research area for detecting the online communities in the available dense social network. The analysis of influence propagation with time series analysis and measuring the centrality between the nodes will significantly increase the metrics of social network. The source node is dedicated to find the required information about the node links. The empirical results were discussed in this paper. Betweenness and centrality of influential node prorogation has identified users in the community and analyzes how the information has exchanged with different time ranges. With unsupervised learning, time ranges are analyzed between 0.5s and 1s in the strong ties of the influential propagation node, and further the influential node relationship between the communities has been calculated by using the strong and weak links. We collect the user information on the online social network-facebook user data, performing some actions post, comments, share, tag, Hit Rate, and the time. From these we have analyzed the propagation between the pair of nodes and users community. It is observed that the influence propagation in the collected data, at a time ’t’ the node U1 which is a ego node of U2, the information was shared between them, if the interaction has performed in a symmetric manner, say at a time $t+\beta$, the action is propagated from U2 to U1. We observed that the interaction of the frequency of action with different content, U2 is indeed exerting the influence on U1. The dependency of discriminate paths was identified and compared with the results obtained from random walk model to deliver a satisfied result by using the proposed model.