<![CDATA[The selection of thesis advisors is a critical step in supporting students’ academic success. This study proposes the hybrid model that combines the C4.5 classification method with the AHP-SAW-TOPSIS multi-criteria decision-making approach to provide objective and systematic advisor recommendations. In the initial phase, the first advisor is selected using four classification algorithms: Decision Tree (C4.5), K-Nearest Neighbor (KNN), Support Vector Machine (SVM) with RBF kernel, and Naive Bayes. The C4.5 algorithm achieved the highest accuracy at 95%. The second advisor is determined using AHP to assign weights to four criteria: supervision load, academic rank, seminar involvement, and research interest alignment. These weights are applied in the SAW method for normalization and initial scoring, followed by TOPSIS to produce the final ranking. The top-ranked advisor is the sixth, followed by the second and tenth. The hybrid approach offers advantages over using either Machine Learning or MCDM alone. Machine Learning excels in identifying patterns from historical data for accurate predictions, while MCDM explicitly incorporates multiple criteria and institutional preferences. Their combination creates a recommendation system that is both precise and policy-aware. Sensitivity analysis shows stable results, indicating that the assigned weights are relevant. This model supports fair, data-driven decision-making and helps reduce the administrative burden of advisor assignments at the university level.]]>
Copyrights © 2025
