Open AccessMethods of Applying Machine Learning to Student Feedback Through Clustering and Sentiment Analysis
Author(s) -
Eric Andersson,
Christopher Dryden,
Chirag Variawa
Publication year - 2018
Publication title -
proceedings of the ... ceea conference
Language(s) - English
Resource type - Journals
ISSN - 2371-5243
DOI - 10.24908/pceea.v0i0.13059
Subject(s) - sentiment analysis , computer science , cluster analysis , relation (database) , artificial intelligence , class (philosophy) , qualitative analysis , key (lock) , online learning , machine learning , exploratory analysis , mathematics education , data science , natural language processing , data mining , qualitative research , multimedia , psychology , sociology , social science , computer security
Machine learning is used to analyze student feedback in first-year engineering courses. This exploratory work builds on previous research at the University of Toronto, where a multi-year investigation used an online survey to collect quantitative and qualitative data from incoming first-year students. [1] (N ~1000)Sentiment analysis, a machine learning method, is used to investigate the relationship between hours of study outside of scheduled instructional hours and qualitative survey feedback sentiment. The results are visualized with chronological sentiment graphs, which contextualize the results in relation to key events during the school year.Large drops in sentiment were seen to occur during weeks with major assessments and deadlines. An inverse correlation between hours spent outside of class and feedback sentiment was also noticed