Abstract
Learning analytics in the post-digital age requires data-driven approaches that go beyond the limits of classical statistical methods. Machine learning (ML) algorithms and large language models (LLM) offer powerful alternatives for complex problems such as predicting student achievement, modeling learning behaviors, and analyzing text-based qualitative data. In this paper, we illustrate the application of ML algorithms such as support vector machines, random forests, and neural networks on training data. We also discuss how NLP and LLMs transform thematic coding processes in the analysis of open-ended responses. It is emphasized that deep learning-based LLMs can successfully capture the themes identified by human researchers, but need human support in contextual nuances. LLMs increase scalability in qualitative analysis with tasks such as summarization, classification and sentiment analysis, and provide researchers with powerful tools to make sense of large data sets. In this context, the integration of ML and LLM-based approaches with traditional analysis techniques paves the way for hybrid solutions that combine explanatory and predictive power in learning analytics.
Details
Presentation Type
Paper Presentation in a Themed Session
Theme
2026 Special Focus—Human-Centered AI Transformations
KEYWORDS
ML Algorithms, LLMs, Learning Analytics
