spStack (development version)
- New model families:
svcLMexact(),svcLMstack(),svcGLMexact()andsvcGLMstack()for spatially varying coefficients model, andstvcLMexact(),stvcLMstack()for Gaussian spatially-temporally varying coefficients model. - Faster, leaner and more robust backend: closed-form exact LOO for the Gaussian models, deletion updates for GLM LOO and K-fold CV, PSIS in C++, informative errors, and a new
diagnosticselement. - Breaking changes: in
spLMexact()/spLMstack()the inverse-gamma prior is on the noise variancesigmaSq; stacked fits storemodel.paramsandstacking.summary. - New lean synthetic datasets
simSpatialandsimSpaceTimereplace the old datasets andsim_spData(); all examples and vignettes use them.surfaceplot()/surfaceplot2()use a colorblind-friendly palette.
spStack 1.1.3
CRAN release: 2026-03-08
- Documentation Update: Pan, Zhang, Bradley and Banerjee (2025) accepted at Bayesian Analysis.
- Migrate to CVXR 1.8.1 API and resolve CRAN Results errors.
- Add a fallback option to the optimization routine.
spStack 1.1.2
CRAN release: 2025-10-04
- Documentation Update: Zhang, Tang and Banerjee (2025) accepted at JASA.
spStack 1.1.1
CRAN release: 2025-07-14
-
lmulm_XTilde_VC(),lmulv_XTilde_VC(): Fixed address sanitizer issue with string comparison with pointer to string literal.
spStack 1.1.0
CRAN release: 2025-07-12
stvcGLMexact(),stvcGLMstack(): New functions for spatially-temporally varying coefficients GLM.posteriorPredict(): New functions for posterior predictive inference using predictive stacking.recoverGLMscale(): Utility for recovering posterior samples of scale parameters in spatial and spatial-temporal GLMs.