Using Natural Language Processing to Facilitate Common Student Misconception Analysis
Author(s)
Zaman, Azreen
DownloadThesis PDF (501.7Kb)
Advisor
Abdelhafez, Mohamed
Terms of use
Metadata
Show full item recordAbstract
There is a large variation in the educational background and purpose of incoming university students. To improve the overall learning experience of these students, we can utilize natural language processing such as topic modeling and sentiment analysis to facilitate common student misconception analysis. This project aims to develop an algorithm via natural language processing that extracts specific topics and common errors that students struggle with in class from online feedback semi-automatically to allow instructors to adjust lesson plans and place emphasis on topics of concern. Using these tools, we can conduct study on the effect on student grades when instructors take into account the information extracted by the model in their lesson plans. This project is aimed at MIT freshmen taking two semesters of physics.
Date issued
2023-06Department
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer SciencePublisher
Massachusetts Institute of Technology