Spatial and spatio-temporal modelling of point patterns using the
log-Gaussian Cox process. Bayesian inference for spatial, spatiotemporal,
multivariate and aggregated point processes using Markov chain Monte Carlo. See Benjamin M. Taylor, Tilman M. Davies, Barry S. Rowlingson, Peter J. Diggle (2015)
CHANGES IN lgcp VERSION 1.2
SIGNIFICANT USER-VISIBLE CHANGES
o Note there is no lgcp version 1.1, although there is a version 1.01, which is apparently the same thing. o The function fftinterpolate.fromXYZ, now uses functions from the raster package to perform interpolation. o Cell inculsion is now by default via grid cells touching the observation window, rather than via centroids.
o Since cell inclusion is now via cells touching the observation window, this should fix the issue of points being `lost due to discretisation', as noted in the package vignette. Users of previous versions of the package should note that the old functionality can be regained by setting "inclusion='centroid'" in the call to lgcpPredict for instance. o There was a bug in the method raster.lgcpgrid, this has now been fixed.
o interpolation of a spatialAtRisk object onto the computational grid is now via functions from the raster package. CHANGES IN lgcp VERSION 1.3-2
o There was a bug in lgcpPredictSpatioTemporalPlusPars if addTemporalCovariates had been used to bolt on temporal-only variables, the columns of the matrix Z needed to be re-ordered to match the order of variables in the overall (spatiotemporal) model formula object. A warning is issued if this happens. CHANGES IN lgcp VERSION 1.3-5
o Added the dependency on rpanel version >= 1.1-3. This fixes the bug with visual parameter estimation caused by changes in the rpanel package. CHANGES IN lgcp VERSION 1.3-15
o Fixed minimum contrast estimation bug caused by updated spatstat functions