Robust Estimation of the Raster-Based Explanatory Model for Land Use Pattern
基于栅格数据的土地利用格局解释模型及应用

The non-linear models with binary variables to identify the existence of some land use types at the pixel-specific dimension are widely used to analyze the relationship between land use pattern and its influential factors. In this paper, we develop and use a Raster-Based Explanatory Model for Land Use Pattern (RBEMLUP) to solve the problems we are always confronted with when we build and estimate the models to figure out the drive mechanism of land use pattern after we introduce the binary variables in the statistical estimation model and when there exist non-liner relationship between the independent variables and dependent variables. The case study in Taips League shows that RBEMLUP is capable of accurately estimating the relationship between the influential factors from the geophysical and social-economic dimensions and the land use pattern at regional extent, through a series of treatments including mainly the multicollinearity diagnoses and singular value elimination. The research results for the case study of Taips League are of significance to supply the decision making information for land use planning and the optional utilization of land resource.