A Framework for Analysing Topics in University Courses
Name
Ragnar Vent
Abstract
The vast amount of courses being taught throughout different faculties of a university makes it difficult to get a general overview of the topics being taught and their overlaps from a singular point of view. The goal of this work is to provide an automated and progressive means to discover such topics, based on the publicly available course materials. The benefits for this kind of analysis range from simply gaining an insight and an overview of the topics covered by different courses, to more specific and goal-oriented purposes. For example, we can discover which courses can support cross-course projects by finding courses that cover the same topic, are taught in the same semester and require the completion of a course project.
The proposed framework is responsible for gathering and extracting course related raw textual content from various heterogeneous sources. The collected data is then cleansed and transformed to a suitable format for further textual analysis, more specifically topic modelling. LDA modelling method is used as a main tool for resolving topics and discovering relations between courses and individual course materials. As a final step of the established mechanism, the analysis results are passed to a predefined visualization component, specifically designed for the framework at hand.
The proposed framework is responsible for gathering and extracting course related raw textual content from various heterogeneous sources. The collected data is then cleansed and transformed to a suitable format for further textual analysis, more specifically topic modelling. LDA modelling method is used as a main tool for resolving topics and discovering relations between courses and individual course materials. As a final step of the established mechanism, the analysis results are passed to a predefined visualization component, specifically designed for the framework at hand.
Graduation Thesis language
English
Graduation Thesis type
Master - Computer Science
Supervisor(s)
Siim Karus
Defence year
2017