Exploring the role of Health Literacy in Mediating the Impact of Social Media Influencers on Students Eating Behaviours

Abstract

The swift emergence of social media influencers has profoundly influenced students' dietary decisions, frequently promoting detrimental eating habits. Nonetheless, the degree to which health literacy can moderate this influence is still inadequately examined. This study examines the behavioral correlation among influencer-driven exposure, health literacy, and students' eating behaviors, introducing an innovative computational-behavioral framework termed SHLRF–IFPA. The model combines the Improved Flower Pollination Algorithm (IFPA) to mimic how influencers affect people and how they filter information, and the Students' Health Literacy Random Forest (SHLRF) classifier to predict behavioral outcomes. The system uses a real-world dataset of 1,108 students to model how social media spreads around the world, how health literacy filters out certain people, how likely it is that mediation will happen, and how people change their behavior. Extensive experimentation shows that the suggested SHLRF–IFPA framework is better than five advanced non-baseline machine-learning classifiers: ExtraTrees, SVM-RBF, MLP-DeepNN, DBN, and AS-QDA. It achieves the best scores across all evaluation metrics, including Accuracy, Precision, Recall, F1 Score, and ROC-AUC (0.982 for each). The results show that health literacy significantly moderates the relationship between influencer exposure and eating behavior, making people less likely to be affected by unhealthy online content. The model also has good scalability and computational efficiency, which makes it useful for monitoring behavior on a large scale. This study provides a powerful predictive instrument and significant insights for educators, policymakers, and health professionals, illustrating that enhancing health literacy can mitigate the adverse dietary effects of social media influencers on students.