Sustainability Evaluation of Rural Development using Geographically Weighted Ridge Regression

In recent scenario, the measuring sustainability of rural development is helpful to identify the targeted strategies and proposing regional differences. The existing regression models struggled with multicollinearity which are sensitive to local variables and computationally expensive. In this research, Geographically Weighted Ridge Regression (GWRR) is proposed for sustainability evaluation of rural development. Initially, the data is collected from National Bureau of Statistics (NBS) where the indicators from the dataset are categorized as economic, environmental, social and cultural respectively. Then, the kernel density estimator is employed to estimate the probability density function for distribution patterns of spatial characteristics to determine the overall dimensional aspects of indicators. After that, Principal Component Analysis (PCA) and weighted entropy are utilized to determine the indicators weights for effective decision-making by reducing dimensionality. In addition, the spatial counties are categorized by determining indicator weights for effective evaluation of sustainability. Finally, the proposed GWRR is introduced to evaluate sustainability by reducing multicollinearity and computational complexity by penalizing the loss and error value. From the results, the proposed GWRR attained outstanding results than existing Geographically Weighted Regression (GWR) with high sustainability rate as 79.93% respectively.