Utilizing Social Network Analysis to Model Social Media Post Popularity with Bayesian Belief Networks

Social Network Analysis (SNA) explores social structures, online interactions, and relationships in social graphs. Using vertices (members) and edges (relationships), SNA represents social networks where members connect through friendships. Platforms like Facebook and Instagram facilitate connection, communication, and information sharing. This paper proposes a novel approach using Bayesian Belief Networks (BBNs) to model social media post popularity. BBNs capture the complex factors influencing popularity, considering followers, engagement, and total likes. This research investigates BBNs' potential to model relationships between user characteristics (followers) and post engagement (engagement rate) in influencing total likes received. The paper proposes a BBN framework with three nodes representing these factors. The methodology defines thresholds to categorize posts into “high” and “low” for followers and engagement rate. Conditional probability tables are then constructed based on these categories, allowing calculation of the probability of a post receiving high likes given its follower base and engagement level. This study demonstrates the feasibility of using BBNs to model social media post popularity. The results reveal that both follower count and engagement rate play a significant role, with a high engagement rate having a stronger influence on total likes, even for posts with a lower follower base.