Search

In recent years, small-area population prediction methods have been widely applied in urban and regional planning. This study aims to empirically compare two small-area population forecasting methods and explains how to choose prediction results through a case study. This study utilizes the geographically weighted regression (GWR) model and the cohort change rate (CCR) method proposed by Inoue to forecast Wuhan's population. In terms of data requirements and the complexity of the prediction process, GWR necessitates considering more drivers of population change and is more effective for population prediction at the county administrative unit or on smaller scales. The model requires substantial input data and involves integrating statistical and geospatial data. Unlike GWR, the CCR method requires population data for only two statistical years, necessitating a smoothing of population data for each area. Consequently, it is generally unsuitable for predicting populations of small geoadministrative units. With regard to forecasting results, the CCR results suggest that Wuhan's future population will continue to experience relatively rapid growth. In contrast, the GWR results indicate that Wuhan's future population will be more stable. Given China's demographic shifts, urbanization trends, and Wuhan's recent population changes, this study concludes that GWR produces results more consistent with reality. The value of this study lies in using a case study to provide researchers and planning practitioners with a better understanding of GWR and CCR and how to choose prediction results for planning. Practical Applications: The study's findings demonstrate that the cohort change rate (CCR) method proposed by Inoue overestimates the actual population, whereas the geographically weighted regression (GWR) method underestimates it. These population prediction methods may produce different results in other cases due to variations in the situations, as observed in Wuhan, China. In the actual planning project research process, multiple prediction methods are often employed to obtain interval prediction results, the median of which is then used to approximate the actual result. To obtain age-specific population prediction results, only the CCR method may be used for planning projects or research. However, the choice of method may be constrained by data availability, requiring results based on the researcher's specific circumstances. [ABSTRACT FROM AUTHOR]