Just to announce this workshop/course: see HERE for the link
which is based on the very exciting INLABRU project!
``EPSRC-sponsored workshop on Spatial Point Process Models for Ecological Survey Data with INLA.
20 – 23 March 2017, University of St Andrews
The Centre for Research into Ecological and Environmental Modelling at the University of St Andrews, in collaboration with the University of Edinburgh, will be holding a 4-day workshop on fitting spatial point processes to data from ecological surveys. The workshop will be held at the custom workshop facility at the Centre for Research into Ecological and Environmental Modelling in St Andrews, Scotland. ...''
Haakon, Elias and Aurelie will be teaching INLA and SPDE. Details here
Aliaksandr Hubin and Geir Storvik (University of Oslo), has just published a report on arxiv comparing the marginal likelihood estimate from INLA with various other estimates; interesting reading...
We congratulate Dr. Gavin Shaddick from Univ of Bath, and colleagues, with all the work with the science behind the new WHO report: Ambient air pollution: A global assessment of exposure and burden of disease which has been all over the news these days. The technical details are described in an arxiv report.
We need to make some backward incomptatible changes for survival models, as several likelihoods might be used in both a regression context and a survival context. Likelihoods that require a inla.surv() object as the reposens, will be named FAMILY.surv, and if the same likelihood is available in a regression context, it will be named FAMILY. The first change is the lognormal, which now exists as 'lognormal' and 'lognormal.surv'. See inla.doc("lognormal") for an example. This change appears in the most recent testing version. Likelihoods that will be affected are: exponential, weibull, loglogistic. If you run weibull survival, then after these future changes, you'll get an error before you change the family name to 'weibull.surv'.
Sorry for the inconvenience.