Achmad Fitro, Nanang Hoesen Hidroes Abbrori, Anita Safitri, Riska Dhenabayu, Renny Sari Dewi, Muhammad Febrilian Dwi Syahputra
Placing students into specialty pathways is an important decision that must consider the balance between student preferences, class quotas, and academic performance during study. This study proposes a framework that integrates the Analytic Hierarchy Process (AHP) method in weighting with a machine learning (ML) model for optimal specialty placement for students, so that it can become a decision support system (DSS). The data are obtained from the university environment, with the final score calculated based on a combination of test results (40 %), student preferences (60 % for the first choice and 30 % for the second choice), and competency scores weighted using AHP. Three classification models, Random Forest, Gradient Boosting, and Neural Network, are trained to predict the most suitable specialty pathway for students' conditions. Among the three models, the Neural Network model shows the highest performance with an accuracy of 97.78 %. The final allocation system considers a maximum quota of 150 students per pathway to ensure decision-making fairness. As a result, 382 students (82 %) were admitted according to their first choice, five students (1 %) were placed in their second choice, and 63 students (1 3 %) were placed in a different track due to quota limitations. This approach demonstrates the effectiveness of integrating AHP and machine learning in supporting fair, transparent, and student-preference-oriented specialty placement decisions. © 2025 IEEE.
Universitas Negeri Surabaya, Digital Business, Faculty of Economics and Business, Indonesia
