Here is some material for my talk and my workshop presentation at the INTECOL 2013 conference.
On Tuesday 20th August at 16.00 (Capital Suite 4) I'll be giving a talk on my new species distribution modelling approach GRaF. Some details can be found here. The slides for my talk can be downloaded from figshare
On Tuesday 20th August at 13.30 (Capital Suite room 7) I'll be showcasing the GRaF R package as part of a workshop on New Computational Methods in R.
The workshop is being put together by [Matthew Smith] (http://research.microsoft.com/en-us/people/mattsmi/) from the BES's computational ecology special interest group which you should join if any of this stuff interests you!
The workshop starts at 12.15, so come along then to see presentations on other new R packages, including Mark Stevenson's work on geographic profiling.
During the workshop, I'll be stepping how to fit models in GRaF and how the approach differs from other SDMs. During the talk I'll go through this tutorial to fit GRaF models to an eel distribution dataset. I'll try and update this with any extra stuff that comes up in the meeting.
If you want to follow along, you can download the R code on its own here
Here are some resources that I found useful for getting to grips with Gaussian random fields (aka Gaussian processes).
Rasmussen & Williams' Gaussian Processes book is free to download and gives a comprehensive overview of GPs. They come from a machine learning perspective and the maths is pretty dense, so it isn't very easy reading.
This blog post shows how to code up your own Gaussian process regression in R and is a bit more accessible.
Part of the aim of the workshop is to discuss and share the process of creating R packages. This is becoming increasingly easy, especially with utilities such as Hadley Wickham's devtools package and functions available in RStudio.
Friedrisch Leisch's tutorial is comprehensive and well worth a read, as is this walk through.
You can download my DPhil thesis over at figshare.
Chapter 5 describes the GRaF species distribution modelling approach which is the subject of the workshop and talk.