CourseGraph course-overlap detection methodology

Research Method · All research methods
Evidence · 3 claims (3 for) · 1 study (1 design), q2 · 0 of 1 report an effect size · 3 claims rest on one study

Description

CourseGraph is a four-step methodology for automatically assessing whether an external course overlaps with courses in a home curriculum. As the authors describe, "CourseGraph first extracts information such as course titles, descriptions, and learning outcomes from a course's webpage using a Large Language Model (LLM) and a tailored prompt." The extracted fields are embedded with sBERT, cosine similarities are computed between course pairs, and size-independent features (max, min, mean similarity, and fractions above 0.5, 0.6, 0.7) feed a 15-dimensional feature vector into a Random Forest classifier that outputs an overlap/no-overlap decision. Embedding each course component separately supports explainability of which components drive an overlap prediction.

Accounts

• CourseGraph: an LLM-plus-sBERT-plus-Random-Forest pipeline for detecting overlap between university courses: CourseGraph is a four-step methodology for automatically assessing whether an external course overlaps with courses in a home curriculum. As the authors describe, "CourseGraph first extracts information such as course titles, descriptions, and learning outcomes from a course's webpage using a Large Language Model (LLM) and a tailored prompt." The extracted fields are embedded with sBERT, cosine similarities are computed between course pairs, and size-independent features (max, min, mean similarity, and fractions above 0.5, 0.6, 0.7) feed a 15-dimensional feature vector into a Random Forest classifier that outputs an overlap/no-overlap decision. Embedding each course component separately supports explainability of which components drive an overlap prediction. (Arthur Nijdam et al. (2026))

Claims

• On the TU/e CS overlap dataset, supervised NLP classifiers (Random Forest, XGBoost) outperform thresholding and zero-shot LLM baselines at detecting course overlap [+W]
• In a worked example from real Erasmus+ decisions, some of CourseGraph's nearest-neighbor course mappings matched the program director's and others did not, one plausibly because it ignores differences in learning methods [+W]
• Component-level embeddings make CourseGraph interpretable: learning outcomes contributed most to an example overlap decision [+W]

Related Research Methods

-

Key Sources

• Arthur Nijdam, Paul Stankovski Wagner, Sara Ramezanian. (2026). CourseGraph: Finding overlaps and differences in Computer Science courses across universities. https://arxiv.org/abs/2608.05910

Page metadata

Field	Value
type	research-method
id	coursegraph-course-overlap-detection-methodology
title	CourseGraph course-overlap detection methodology
description	CourseGraph is a four-step methodology for automatically assessing whether an external course overlaps with courses in a home curriculum.
status	review
generated	by process:settle-candidates on 2026-10-10
trust	unverified — no one has confirmed this page's content

Frontmatter is the page's machine-readable record — see the Schema & Guide.