Predicting Social Media Engagement in Higher Educational Institutions: A Machine Learning Approach

Abstract

Social media engagement plays a vital part in the marketing activities of Higher Educational Institutes (HEIs). The extant studies clearly demonstrate that having a social media presence and posting content alone without a strategy in place would not lead to higher awareness, branding, or enrolments of HEIs. Accordingly, this study, attempts to build predictive models for social media engagement for HEIs in an emerging market context. The official Facebook insights data of a HEI have been utilized for this model-building purpose. Pre-processing and feature engineering steps are performed initially, followed by cluster-based-class-labelling. The labelled classes are further trained with several machine learning models, and the best-performing ones are selected based on K-fold cross-validation. Hyperparameter optimization has been further performed on the best-chosen classifiers. This study is a pioneering attempt to build a predictive model for social media engagements of HEIs in an emerging market context. The proposed predictive models would be a handy tool for devising social media content posting strategies.