Management of sustainable urban green spaces through machine learning-supported MCDM and GIS integration

This study evaluates green space suitability in İzmir's Konak district using the analytic hierarchy process, machine learning, weighted linear combination, and the technique for order preference by similarity to ideal solution methods, integrated with geographic information systems. The approach enhances reliability in green space identification by ensuring consistent integration of weights determined by different methods. Machine learning enables dynamic adjustments to criterion weights, yielding the best results with the random forest algorithm. The analysis revealed that 75% of green spaces were sub-optimally located, with optimal zones in the western and northern regions. The technique for order preference by similarity to ideal solution methodology prioritized area ZGS_6 as the most suitable area, while ZGS_4 ranked lowest. This method supports efficient resource allocation and improves budgeting processes. Hence, the integration of multi-criteria decision-making and machine learning with geographic information systems enhances the planning of sustainable cities and offers critical insights to decision-makers who prioritize sustainability and livability.