Epigraphiology: A Hybrid Approach for Measuring and Analyzing Influence Diffusion in Article Networks

Identifying influential nodes in an article network is crucial for understanding the dynamics of information propagation and its impact on various applications. Traditional methods often rely on citation-based analysis or network structure, overlooking the intricate dynamics of diffusion and node linkages. In this research, we propose a novel scoring model, named “Epigraphiology,” which combines these aspects to compute and analyze the elements contributing to the spread of influence in article networks. To evaluate the effectiveness of our approach, we employ real published article networks with around 904 articles downloaded from the WOS (Web of Science) with total cited references of 32084 in the field of cloud computing from 2010 to 2015. By leveraging the SIR (Susceptible-Infected-Removed) model, we compare the dynamics of articles in the network with the transition of states, highlighting the diffusion process. Additionally, we derive the Reproduction number (R0) for our model, serving as an indicator of the potential spread of influence. Our findings showcase the following key contributions: (a) Epigraphiology introduces a novel methodology for measuring the diffusion capacity of an article’s influence in a hybrid manner, combining diffusion dynamics and node linkages. (b) Contrary to traditional approaches that primarily consider the number of citations (in degree), our results reveal that articles with lower citation counts can still act as super-spreaders, reflecting the ground-truth influence scores. Cross-validation of an article’s influence diffusion score is performed, shedding light on the significant factors contributing to its spread within the network. By bridging the gap between diffusion dynamics, node linkages, and influence measurement, Epigraphiology offers a comprehensive approach to understanding and quantifying the spread of influence in article networks. This research holds implications for various fields and applications where the identification of influential spreaders is paramount in leveraging information dissemination and impact assessment.