A Hybrid Machine Learning Framework for Measuring Social Media Marketing Effectiveness in Higher Education Institutions

Abstract

The swift expansion of social media has revolutionized the marketing approaches of Higher Education Institutions (HEIs), allowing universities to enhance their institutional branding, connect with potential students, and facilitate communication with various stakeholders. Platforms such as Facebook, Instagram, LinkedIn, X (formerly Twitter), and YouTube have emerged as vital channels for student recruitment and digital outreach. Nevertheless, assessing the effectiveness of social media marketing poses a considerable challenge, as traditional evaluation methods predominantly depend on descriptive metrics like likes, shares, comments, impressions, and follower counts. These metrics frequently do not adequately reflect the intricate relationships between user engagement, institutional reputation, and marketing results, underscoring the necessity for more sophisticated and data-driven evaluation techniques. This research introduces a Hybrid Machine Learning (HML) Framework designed to assess the effectiveness of social media marketing within Higher Education Institutions. The framework incorporates advanced data preprocessing, feature engineering, feature selection, ensemble learning, and Explainable Artificial Intelligence (XAI) to enhance predictive accuracy and model interpretability. Essential engagement features, such as reach, impressions, engagement rate, click-through rate, comments, shares, follower growth, video views, and sentiment scores, are utilized to assess campaign performance. The proposed framework leverages the complementary strengths of Random Forest, XGBoost, LightGBM, and Support Vector Machine (SVM) through an ensemble learning strategy, facilitating robust predictions across diverse social media datasets. Moreover, the framework integrates SHAP (Shapley Additive Explanations) to pinpoint key engagement factors and deliver clear explanations for model predictions. This improvement facilitates decision-making by allowing university administrators and digital marketing experts to refine content strategies, enhance student recruitment initiatives, and allocate marketing resources more efficiently. The proposed framework advances educational marketing analytics by offering a scalable, interpretable, and data-driven solution that addresses the shortcomings of single-model methodologies. It provides actionable decision support for assessing and optimizing social media marketing effectiveness, thereby aiding the digital transformation and sustainable development of Higher Education Institutions.