Mining Student Feedback to Improve the Quality of Higher Education through Multi Label Classification, Sentiment Analysis, and Trend Topic

摘要

This research carried out the label aspect classification, sentiment analysis, and topic trends on the Open-Ended Question (OEQ) section for Student Feedback Questionnaire (SFQ). Multi-Class aspect label classification for SFQ will choose the best classification model by comparing the results of the evaluation of accuracy, precision, recall, and Flscore for each feature combination and comparison of four classification algorithms namely Decision Tree (DT), Naive Bayes (NB), K-Nearest Neighbor (KNN), and Support Vector Machine (SVM). The results of this research are Classification Techniques using a combination of features of TFIDF, Unigranb and Bigram with the SVM algorithm which is the best Multi-Class classification model for labeling SFQ aspects. In addition, the SentiStrenghtID algorithm used to get sentiments and also the LDA (Latent Dirichlet Allocation) used to get annual topic trends on each survey aspect label. The findings can help Higher Education to support decision making in taking proactive actions towards improvement for self-evaluation and quality.