Using sentiment analysis to evaluate qualitative students' responses

Area of Science:

• Educational Technology
• Natural Language Processing
• Data Science

Background:

• Text analytics is integral to SMART campus development.
• Sentiment analysis of qualitative student feedback is vital for educational institutions.
• Understanding learner emotions and mood during class engagements is crucial.

Purpose of the Study:

• To evaluate the effectiveness of text analytics, specifically sentiment analysis, in processing qualitative student feedback.
• To compare the performance of four machine learning classifiers in classifying student sentiment.

Main Methods:

• Deployed Naïve Bayes (NB), Support Vector Machine (SVM), J48 Decision Tree (DT), and Random Forest (RF) classifiers.
• Utilized qualitative feedback text from a semester-based course at the University of Education, Winneba.
• Employed k-fold cross-validation for training and testing classifier models.

Main Results:

• The Support Vector Machine (SVM) classification algorithm demonstrated superior performance.
• SVM achieved the highest accuracy rate of 63.79% in sentiment classification.
• Other classifiers like NB, DT, and RF were also evaluated.

Conclusions:

• Sentiment analysis using text analytics is a viable tool for understanding student feedback in higher education.
• SVM shows promise as an effective classifier for analyzing qualitative student sentiment data.
• The findings contribute to the advancement of SMART campus architectures through data-driven insights.