Research Information System
Applied Spatial Analysis and Policy, vol.18, no.4, 2025 (SSCI, Scopus)
This study presents an innovative approach that integrates Geographic Information Systems (GIS) and Machine Learning (ML) techniques to optimize school site selection in the Konak and Karabağlar districts of İzmir. The study was conducted using 12 different criteria and 39 alternative sites by integrating the Analytic Hierarchy Process (AHP), Equal Weighting, and Extreme Gradient Boosting (XGBoost) methods with the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). The alternative sites were analyzed based on the average pixel values of the criteria, and AS4 was identified as the most suitable site. The ML-based method, ML_Rank, significantly contributed to the ranking of alternatives, showing high compatibility with the Equal Weighting and XGBoost methods (ρ = 0.9745 and ρ = 0.8813, respectively). In contrast, AHP exhibited the lowest correlation among the methods, highlighting the superior objectivity and consistency of ML-based approaches in ranking alternatives. The scalable and automated structure of ML_Rank enabled the rapid and reliable analysis of large datasets and complex criteria. Spearman correlation analysis demonstrated that ML-based methods improve decision-making processes by producing objective and consistent results. Thus, the findings reveal that GIS and ML techniques can be effectively utilized in critical planning processes such as school site selection. These methods contribute to sustainable urban planning by supporting the equitable distribution of educational services, enhancing student accessibility, and promoting efficient resource utilization.