Simpson 2016 - Going Off Grid LGCP Inference

Routing Summary

Full capture of Simpson, Illian, Lindgren, Sorbye and Rue (2016), Going off grid: Computationally efficient inference for log-Gaussian Cox processes (Biometrika; arXiv:1111.0641), plus a concept note on Gaussian random fields and kriging (in the cluster folder Spatial Fields Concept Notes).

Concept Map

ConceptNoteTypeDepends OnKey Result
IPP and LGCP likelihood; lattice approximationSimpson 2016 - LGCP Likelihood and the Lattice Approximationdefinitionnonebinning is the dominant lattice error
Off-grid Poisson-form likelihoodSimpson 2016 - Off-Grid Poisson-Form LikelihoodtheoremLGCPexact point term + quadrature integral
Matern SPDE, FEM weightsSimpson 2016 - Matern SPDE and GMRF Representationdefinitionoff-grid
ConvergenceSimpson 2016 - Convergence Theorytheorembothlattice vs
Case studiesSimpson 2016 - Examples BCI Sampling Effort and Oceanexampleall37 s to 24 s with coarsened mesh
GRF, kriging, field-level inferenceSpatial Fields - Gaussian Random Fields, Kriging and Field-Level InferenceconceptSPDEkriging mean/variance; spectral diagonalisation

Notes

Sources

Simpson 2016 - Going Off Grid LGCP Inference (arXiv:1111.0641v3). LGCP original: Moller, Syversveen & Waagepetersen (1998, Scand. J. Stat.); no legal free copy was retrievable, covered here via Simpson et al.

See Also

Hilbert Space Gaussian Processes | Gaussian Process Regression | Spatial Models - BYM