/ en / Traditional / mobile

Beats Biblionetz - Texte

What Are We Teaching?

Automated Evaluation of CS Curricula Content Using Topic Modeling
Jean Michel Rouly, Huzefa Rangwala, Aditya Johri
Publikationsdatum:
Zu finden in: ICER 2015 (Seite 189 bis 197), 2015
Erste Seite des Textes (PDF-Thumbnail)

iconZusammenfassungen

Identifying the concepts covered in a university course based on a high level description is a necessary step in the evaluation of a university's program of study. To this end, data describing university courses is readily available on the Internet in vast quantities. However, understanding natural language course descriptions requires manual inspection and, often, implicit knowledge of the subject area. Additionally, a holistic approach to curricular evaluation involves analysis of the prerequisite structure within a department, specifically the conceptual overlap between courses in a prerequisite chain. In this work we apply existing topic modeling techniques to sets of course descriptions extracted from publicly available university course catalogs. The inferred topic models correspond to concepts taught in the described courses. The inference process is unsupervised and generates topics without the need for manual inspection. We present an application framework for data ingestion and processing, along with a user-facing web-based application for inferred topic presentation. The software provides tools to view the inferred topics for a university's courses, quickly compare departments by their topic composition, and visually analyze conceptual overlap in departmental prerequisite structures. The tool is available online at http://edmine.cs.gmu.edu.

Von Jean Michel Rouly, Huzefa Rangwala, Aditya Johri im Konferenz-Band ICER 2015 im Text What Are We Teaching? (2015)

iconAnderswo finden

icon

iconVolltext dieses Dokuments

LokalAuf dem WWW What Are We Teaching?: Fulltext at the ACM Digital Library (lokal: PDF, 100 kByte; WWW: Link OK 2017-06-28)

iconAnderswo suchen Auch im Biblionetz finden Sie nicht alles. Aus diesem Grund bietet das Biblionetz bereits ausgefüllte Suchformulare für verschiedene Suchdienste an. Biblionetztreffer werden dabei ausgeschlossen.

iconBiblionetz-History Dies ist eine graphische Darstellung, wann wie viele Verweise von und zu diesem Objekt ins Biblionetz eingetragen wurden und wie oft die Seite abgerufen wurde.

Verweise auf dieses Konferenz-Paper 1
Verweise von diesem Konferenz-Paper 1
Webzugriffe auf dieses Konferenz-Paper 331252732
20162017