News

Preview of new book: Advanced Spatial Modeling with Stochastic Partial Differential Equations Using R and INLA

posted 22 Sep 2018, 02:02 by Havard Rue

This is to announce a forthcoming new book: Advanced Spatial Modeling with Stochastic Partial Differential Equations Using R and INLA, by Elias T. Krainski, Virgilio Gómez-Rubio, Haakon Bakka, Amanda Lenzi, Daniela Castro-Camilo, Daniel Simpson, Finn Lindgren and Håvard Rue, to be published by CRC Press/Taylor and Francis Group, 2019.

For more details, see here, which also includes a link to a (free)  Github-version of the book.

H

Space-time INLA workshop

posted 6 Sep 2018, 03:15 by Haakon Bakka

We would like to announce a workshop on SPDE modeling and INLA, which will take place in the South of France in Avignon from November 7 to 9. The workshop will feature tutorials on INLA/SPDE as well as scientific talks on topics related to the INLA/SPDE universe. It is organized by INRA's RESSTE network ("space-time modeling") and the French EcoStat network ("statistical ecology").

Our invited speakers (some of them also giving practical tutorials) are:

Haakon Bakka, KAUST
David Bolin, Chalmers University
Nicolas Desassis, MinesParisTech
Janine Illian, University of St Andrews
Finn Lindgren, University of Edinburgh


We will soon provide a more detailed program.

We will also schedule a number of contributed talks involving SPDE approaches. If interested, please provide a title and a short summary. We might also open a poster session.

The program of the workshop starts on Wednesday, November 7 at 9am and will finish November 8 at the end of the afternoon. Friday November 9 will be reserved for free discussions, for revisiting some of the concepts discussed during the preceding days, and it will also be possible to work on the participants' data if they wish.

Registration is free but mandatory, within the limit of available places.
In case you would like to participate, please send me a message (thomas POINT opitz AT inra POINT fr) indicating your name, affiliation and status (PhD, postdoc, Professor, ...).


The official language of the workshop will be English (even if you may have seen some announcements in French).

We would be happy to welcome you in Avignon soon!


Best regards

The organizers (Denis Allard, Edith Gabriel, Thomas Opitz)

Forwarded by Haakon Bakka


Want to do a PhD?

posted 1 Sep 2018, 10:05 by Havard Rue

I have an opening for a new PhD student in my group here at KAUST. Application deadline is about mid October.  Contact me if your are interested.

H

Missing links to be expected...

posted 1 Sep 2018, 07:01 by Havard Rue

Due to my move to KAUST last year, I had to close down the www-hosting of some of the old files within this site that were stored at server at NTNU. New files have been added for quite some time with the new links to inla.r-inla-download.org, but some old links still remain.  If you get a missing link of the type

    http://www.math.ntnu.no/inla/r-inla.org/case-studies/muff-etal-2014/muff-etal-2014-data-code.zip

the translation to the new link, should be

    http://inla.r-inla-download.org/r-inla.org/case-studies/muff-etal-2014/muff-etal-2014-data-code.zip

where the required change is shown in boldface. 

All these links have to be edited manually, so I'll do those I find.  If we miss someone, please let us know at help@r-inla.org

H




Alternative Linux builds

posted 23 Jul 2018, 08:04 by Havard Rue   [ updated 24 Jul 2018, 09:42 ]

The R-INLA package comes with Linux binaries built on Ubuntu1604. There are some alternative Linux builds available at 

http://inla.r-inla-download.org/Linux-builds/


You can install one of these, by replacing the directory '64bit' at 

> dirname(INLA:::inla.call.builtin())

which gives for me

[1] "/home/hrue/R/x86_64-redhat-linux-gnu-library/3.5/INLA/bin/linux/64bit"


with the one you download. Alternatively, you can store it somewhere else and set variables 'inla.call' and 'fmesher.call', using 'inla.setOption', to the full path to 'inla.run' and 'fmesher.run' in that directory.

Static builds are available, doing

        INLA:::inla.dynload.workaround()

The static build is without 'rgeneric' support and is also built on Ubuntu1604.


Spatial modelling with R-INLA workshop: 18-20 Sept @Uni-Sheffield

posted 23 Jul 2018, 07:55 by Havard Rue



***********************

DESCRIPTION

Are you interested in learning state-of-the-art methods in spatial modelling?

We are running a 3-day workshop on applications of R-INLA http://www.r-inla.org/ to environmental data. The workshop will take place on 18-20 September 2018 at the University of Sheffield.

The workshop aims to bridge the gap between the environmental science and geostatistical community.

No prior assumed knowledge of GIS or R are required. The workshop is open to academic staff and postgraduate research students on a first-come basis. 50 places are available.


R version 3.5

posted 12 Jul 2018, 04:27 by Havard Rue

We have moved to R-3.5. There is a testing/stable version of INLA Version_18.07.11  that is built with 3.4, and essentially the same testing/stable version is also available as Version_18.07.12 built with R-3.5.  

You have to reinstall all packages when moving from R-3.4 to R-3.5, due to an internal change in the binary format. 

Please let us know if there are any problems with INLA & R-3.5.

H

Upcoming INLA-related tutorials at useR! and RaukR

posted 11 Jun 2018, 01:22 by Havard Rue

Paula Moraga from Lancaster Medical school, will give two INLA related tutorials at the upcoming useR! meeting in Brisbane and the RaukR summer school in Visby, Sweden. The link to the tutorials (and other useful stuff!) is here

H



New web address for shiny application SSTCDapp

posted 27 May 2018, 13:37 by Andrea Riebler

The shiny application SSTCDapp developed by the spatial statistics group at the Public University of Navarre moved to a new location




SSTCDapp has been designed for the following purposes:

  • To perform descriptive analyses in space and time of mortality/incidence risks or rates.
  • To fit a wide variety of spatial and spatio-temporal hierarchical models commonly used in disease mapping.

The SSTCDapp facilitates the use of these models, since some of them are difficult to implement for non-expert users working in epidemiology or public health institutes. The application may also be used for the analysis of similar problems in many other fields such as criminology, gender-based violence, road traffic accidents or veterinary epidemiology.

The application allows the users to upload their data and its associated cartography, to generate many graphs and tables for the descriptive analysis of mortality/incidence of risks/rates, and last but not least, to fit fairly complex spatio-temporal models to smooth risks in space and time. The recently proposed integrated nested Laplace approximation (INLA) technique for Bayesian inference is used for model fitting and inference using the R-INLA package.

How besag/bym and those models should be defined...

posted 23 May 2018, 12:07 by Havard Rue   [ updated 23 May 2018, 12:17 ]

The besag/bym model has been around for quite some time, and is extensively used.  When applied to disconnected graphs, the current practice has varied, as the original formulation by BYM, implicitly assumed a connected graph.  Anna.FS & Massimo.V  have now defined this in general for a disconnected graph, including how to do proper scaling of the model so that priors make sense. This is how besag/bym/etc with scale.model=TRUE is implemented in R-INLA. The link to the paper is here, and is free before 12th July 2018. Otherwise, you'll find older versions on arxiv.

H



1-10 of 134