THE USE OF MACHINE LEARNING TO IDENTIFY SUITABLE AREAS FOR URBAN GROWTH IN MOUNTAINOUS AREAS: TUNCELİ CITY EXAMPLE

Abstract

One of the most important trigger factors contributing to increased human intervention in space in many regions of the world is urbanization. To manage and plan urbanization in harmony with other human activities, it is necessary to manage and plan it accordingly. Even though urbanization studies tend to focus on large cities, small-scale cities are quite common throughout the world, both in terms of their numbers and regarding their population density. Moreover, small cities can contribute to a more homogeneous distribution of development at the national and regional levels. It may, however, be hindered by a variety of limitations, including the hinterlands and the unused potential of these settlements. The city of Tunceli is also a small settlement with natural and human factors limiting its growth. In this study, based on machine learning algorithms, “support vector machines”, “artificial neural networks” and “random forest” models were used to determine urban growth zones. In the city, the most suitable sites for primary growth are those which are suited for peripheral growth and inward-stacked growth (12 km2). While more than 90% of predictions were accurate, regarding the spatial equivalents of the findings, the best results respectively, came from “random forests”, “artificial neural networks”, and finally “support vector machines”.