International Journal of
Information and Education Technology

Abstract—Sentiment analysis has emerged as a complementary aid to opinion extraction in textual information like academic faculty evaluations. Selecting the best machine learning algorithm involves balancing accuracy and computational complexity. This work compares classical machine learning algorithms, including Logistic Regression, Support Vector Machine (SVM), Naïve Bayes, Random Forest, and k-Nearest Neighbors (k-NN), for classifying student evaluations into negative, neutral, and positive sentiment. Training and testing were performed on labeled data, and performance was evaluated using precision, recall, F1-score, overall accuracy, a confusion matrix, and computational time. Random Forest was most accurate (99%) but took maximum computational time (4.575s). SVM came in a close second at 97%, with negligible computational cost, making it a formidable contender in applications where efficiency is a priority. Logistic Regression came in next at 94%, while Naïve Bayes was quickest (0.019s) but slightly less accurate (93%). k-NN performed worst at 88% and struggled to discriminate between overlapping sentiment classes. Confusion matrix analysis pointed out that Random Forest and SVM reduced misclassification primarily between neutral and positive sentiments. Analysis points out a trade-off between predictive accuracy and computational efficiency. Random Forest is superior for high-stakes sentiment classification, where predictive accuracy is prioritized above all else. Naïve Bayes and SVMs are preferred choices for applications that require real-time responses or where computational power is a limiting factor. Additional research should incorporate deep learning architectures and broader efficiency metrics, such as memory usage and energy consumption, to solidify applicable use cases for sentiment classification in education.

Keywords—computational efficiency, faculty evaluations, machine learning, Random Forest, sentiment analysis, Support Vector Machine (SVM), text classification

[PDF]

Cite: Aaron Paul M. Dela Rosa, Renato L. Adriano II, and Lilibeth G. Antonio, "Benchmarking Classical Machine Learning Models for Sentiment Classification in Students’ Faculty Evaluations: A Comparative Study," International Journal of Information and Education Technology, vol. 16, no. 7, pp. 1844-1855, 2026.

Copyright © 2026 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).