Methods for calculation the spatial distribution of the territory membership to the urbanized based on the hybrid neural network

Abstract

This paper presents the method for calculation of the spatial distribution of the territory membership to the urbanized (TMU)—the value that puts the degree of suitability for development (tourism development) in accordance with each node of the grid of the study area within (0, 1). The method is based on a hybrid neural network training. It allows us to assess all kinds of territories for tourist infrastructure development capabilities with relation to attractiveness and provide results visualization on the geographical information system (GIS) maps. The developed algorithm of picking and conversion of geospatial data from GIS for knowledge base generation uses attractors coordinates (vectors of roads, a city center, a railway station) and random points of the explored area, makes possible distance calculation between them (on the road to the attractor and Manhattan distance) and following conversion to the ASCII file, that allows unifying input parameters of the set of models for forecasting development of tourist infrastructure objects. The paper studied typical tourist towns of the Ukrainian Carpathians.