<![CDATA[Choosing a thesis topic remains challenging because many students struggle to select an appropriate research title. Thesis title classification can support this process by helping identify suitable research areas. Although classification algorithms have been widely applied in sentiment analysis, comparative studies of Random Forest, Support Vector Machine (SVM), and Naive Bayes, for thesis title classification using N-Gram features remain limited. This study compares these algorithms through text preprocessing, TF-IDF-based N-Gram feature extraction, and evaluation using confusion matrices and processing time. The dataset consists of 96 thesis titles classified into four categories: Information Systems, Multimedia, Networks, and IoT & Artificial Intelligence. The results show that SVM achieved the best performance, with 80% accuracy, 89% precision, 80% recall, an F1-score of 81.46%, and a processing time of 0.003 seconds, indicating that SVM is the most effective algorithm for thesis title classification compared with Naive Bayes and Random Forest.]]>
Copyrights © 2026
