A Spatially Heterogeneous Expert Based (SHEB) Urban Growth Model Using Model Regionalization

摘要: Urbanization changes have been widely examined and numerous urban growth models proposed. We introduce an alternative model specifically designed to incorporate spatial heterogeneity in models. Instead of applying a single method the entire study area, we segment area into different regions apply targeted algorithms each subregion. The working hypothesis is that integration appropriately selected region-specific will outperform globally applied as it further heterogeneity. examine land use Denver, Colorado. Two maps from time snapshots (1977 1997) are used detect changes, 23 explanatory factors produced urbanization. proposed Spatially Heterogeneous Expert Based (SHEB) tested decision trees underlying modeling algorithm, them subregions. In this paper segmentation division interior exterior areas. Interior areas those situated within dense urbanized structures, while outside these structures. Obtained results on regionalization technique indicate local produce improved terms Kappa, accuracy percentage multi-scale performance. superiority also confirmed by pairwise comparisons using t-tests. criterion interior/exterior selection may not only capture specific characteristics morphological properties, but socioeconomic which implicitly be present representations. usage subregions acts proof concept. Other indicators, for example landscape, political boundaries could act basis segmentations.