Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore a conference presentation that introduces an innovative approach to personalizing programming education through automated exercise classification. Learn how researchers from Eindhoven University of Technology combine lexicon-based analysis with machine learning and natural language processing techniques to classify programming exercises by both topic and difficulty level. Discover the methodology behind using BERTopic for topic modeling and five different machine learning models to predict exercise difficulty levels, based on a dataset of 106 programming exercise descriptions from introductory courses and performance data from up to 189 learners. Understand how lexicon-based approaches significantly improve topic modeling accuracy and coherence compared to baseline methods, while also providing modest but consistent gains in difficulty prediction despite the inherent challenges of defining ground truth for exercise complexity. Gain insights into the foundational work that enables scalable and resource-efficient solutions for generating personalized programming exercises, addressing the growing need for automated content generation in programming education.
