Integrating Rule-based Decision-Making and Predictive Modeling for Enhanced user Engagement in Social Networks

This paper explores a rule-based decision-making approach to guide user behavior on social network platforms. By applying predefined rules based on user activity data, such as likes, followers, posts, and videos, personalized recommendations are generated to improve engagement and content interaction. The thresholds for these rules are determined using statistical methods, specifically the median of user data attributes. Predictive models, encompassing Logistic Regression (LR), Random Forest (RF), as well as Gradient Boosting, complement rule-based approach to identify high-engagement users. Results demonstrate that complex models such as RF, along with Gradient Boosting, outperform simpler models in predicting user engagement. This framework provides a structured way to analyze and enhance engagement strategies on social media platforms. The novelty of this work is the combination of a rule-based decision system with machine learning models for user engagement prediction. Most existing studies mainly focus on prediction only. The proposed framework not only predicts user engagement but also provides useful recommendations for improving user activity and interaction.