Publication Details
Abstract
The objective of the paper is to compare the SLRM with the WM and highlights the practical applications of both models in the spatial data analysis. Spatial linear regression is an important statistical method for spatial dependence data analysis, which can be used for spatial pattern detection and variation exploration of the field. In comparison, the weighted model is a way of estimating data by borrowing the weights of the relationships between points, for focused on the level of examining the differences among points (locations). Both models were implemented on spatial datasets to make predictions, this study only aims at comparing the accuracy of prediction for the two models on spatial data using statistical indicators mean absolute error (MAE) and root mean square error (RMSE). The findings suggest that there are pros and cons of both models, and the model choice depends on the properties of the data under study, as well as the intensity of the spatial relationship across observations. The results of this study suggest that the weighted model might be more applicable in regions with great fluctuation and inhomogeneous characteristics between sites, and the SLR model can be more applicable in the study of the continuously correlated spatial relationships.