A methodology for surveying gradients of influence on social media platforms using multi-media content

Determining the importance of influence inside a network and its impact on the behaviour of its users can provide insight into historical trends, information dispersal, and reducing the spread of misinformation. Therefore, improving research into how users perceive and interact with others is valuable. Understanding how members of a group influence each other by sharing media or holding conversations can be especially vital. These interactions (often between two group members) can lead to members adopting certain behaviours, we consider this an example of influence. In this paper, we review existing work in detecting and defining influence in social networks and we propose a methodology for three experiments using content features and transformer architecture models to measure and evaluate different types of influence at various resolutions