Enhancing Online Education Through Sentiment Analysis and Complex Systems Modelling

This research explores the application of sentiment analysis through the lens of complex systems modelling to enhance the quality of online certification courses, with a particular focus on global platforms such as Coursera. The COVID-19 pandemic catalyzed significant growth in online learning, creating an urgent need for adaptive and student-centric approaches to ensure relevance and effectiveness. Leveraging unstructured textual data from student reviews of courses, this study integrates methodologies from systems science, computer science, and education to address real-world challenges in online education. By employing both lexicon-based (SentiWordNet and VADER) and supervised machine learning techniques (Multinomial Naive Bayes, Support Vector Machine, and Stochastic Gradient Descent), the research conducts a detailed sentiment analysis to identify patterns, emergent behaviours, and feedback loops inherent in course design and delivery. Findings reveal that Support Vector Machine achieves the highest accuracy at 97.3%, offering insights that guide iterative improvements in course content and pedagogical strategies. The study demonstrates how interdisciplinary approaches to sentiment analysis can inform responsive education environments, aligning with broader societal goals of accessibility, inclusivity, and quality in online learning ecosystems.