Spatial–temporal correlation dimension reduction and unsteady granger driving factor in regional varied earth surface in GIS

Abstract Spatial–temporal relationships are fundamental. This paper focuses on the improvement for relationship redundancy and causality fuzzy by regional varied earth surface. We propose a method of spatial–temporal correlation dimension reduction and unsteady Granger driving factor in regional var...
Ausführliche Beschreibung

Abstract Spatial–temporal relationships are fundamental. This paper focuses on the improvement for relationship redundancy and causality fuzzy by regional varied earth surface. We propose a method of spatial–temporal correlation dimension reduction and unsteady Granger driving factor in regional varied earth surface. Clustering of hybrid spatiotemporal correlation based on manifold is further proposed to solve the problems of relationship redundancy. Nonlinear Granger causality is introduced to analyze the driving factor for regional varied earth surface. Through the simulation and the actual data analysis, the new method proposed in this paper can improve the residual error estimation of 20.5% and the mean square error of 36.2%. Ausführliche Beschreibung