Analyzing students’ opinions on teaching and learning management using data mining techniques

Abstract

In universities, the management of teaching and learning plays a crucial role in student learning. Gathering student
feedback is important for improving the quality of teaching and learning, allowing instructors to adjust teaching methods, content, or activities to better suit students. In courses with a large number of students, summarizing all feedback for the purpose of improving teaching can be time-consuming. Therefore, this research aims to develop a model for analyzing student opinions on teaching and learning management using data mining techniques. The research methodology is divided into five steps: 1) Data collection of 3,000 student feedback messages. 2) Data preparation by filtering messages and performing word segmentation, along with selecting key terms using the TF-IDF technique. 3) Modeling to compare the performance of four algorithms: Support Vector Machine (SVM), K-Nearest Neighbors (K-NN), Naive Bayes, and Random Forest for classifying positive and negative opinions and categorizing feedback on various aspects such as instructors, course content, and learning support resources. 4) Model performance evaluation using accuracy, precision, recall, and F1-score, and 5) Model application. The research findings indicate that the SVM algorithm has the highest performance in classifying positive and negative opinions, with average values of accuracy, precision, recall, and F1-score at 97.00%, 97.10%, 97.40%, and 97.30%, respectively. For categorizing feedback on various aspects, the KNN algorithm demonstrated the best performance, with average values of accuracy, precision, recall, and F1-score at 91.00%, 91.60%, 91.00%, and 91.10, respectively. The developed model has been deployed as a web application to analyze student feedback, effectively enhancing the quality of teaching and learning management.