Correlation analysis between crown width and DBH using Geographically Weighted Regression

As for simulating Crown-Width Model, the traditional linear models often disregard the spatial component of data. However, estimating a local model reveals that the influence of the independent variables is inconstant over the whole study area. In this research, classic linear regression was reviewed, and Geographically Weighted Regression (GWR) method which belongs to local estimation was introduced. The GWR method is able to show local patterns in the spatial distribution of a parameter, which would be ignored by the OLS approach. The relationship Between Crown Width and DBH was discussed as a case study. After analyzing attribute and spatial information of the 60 trees, it is clear that there was a significant improvement in the Crown-Width model performance of GWR over ordinary least square (OLS).