STWR
Spatiotemporal Weighted Regression (STWR)
Spatiotemporal Weighted Regression Model for Analyzing Local Non-stationarity in Space and Time
Github
https://github.com/quexiang/STWR
Background
This project/algorithm is developed by collaborating with
Xiang Que (Search Results Web results Fujian Agriculture and Forestry University),
Xiaogang Ma (University of Idaho)
Qiyu Chen (China University of Geosciences-Wuhan)
News
The paper has been published: Que, X., Ma, X., Ma, C.*, and Chen, Q., 2020. A spatiotemporal weighted regression model (STWR v1.0) for analyzing local nonstationarity in space and time, Geosci. Model Dev., 13, 6149–6164, https://doi.org/10.5194/gmd-13-6149-2020.
Spatiotemporal Weighted Regression
Features
STWR model calibration via a new spatiotemporal kernel. And it can use data observed at different past time stages to make the model better fit the latest observation points. A highlight of STWR is a new temporal kernel function, in which the method for temporal weighting is based on the degree of impact from each observed point to a regression point. The degree of impact, in turn, is based on the rate of value variation of the nearby observed point during the time interval. The updated spatiotemporal kernel function is based on a weighted combination of the temporal kernel with a commonly used spatial kernel (Gaussian or bi-square) by specifying a linear function of spatial bandwidth versus time.
GWR model calibration via iteratively weighted least squares for Gaussian, Poisson, and binomial probability models.
GWR bandwidth selection via golden section search or equal interval search
GWR-specific model diagnostics, including a multiple hypothesis test correction and local collinearity
Monte Carlo test for spatial variability of parameter estimate surfaces
GWR-based spatial prediction
MGWR model calibration via GAM iterative backfitting for Gaussian model
MGWR covariate-specific inference, including a multiple hypothesis test correction and local collinearity