What this book is and isn’t
This book has a somewhat unusual format, and we need to clarify this to avoid confusing readers. The goal of this book is to provide several tools for cutting-edge research in applied spatial modeling, in under \(300\) pages. To achieve this goal we made a few unusual decisions when writing this book:
No detailed introduction: there is no detailed introduction to spatial modeling or spatial processes in this book. We assume the reader is familiar with this literature or has other sources to depend on.
The explanations of the models are limited to the minimum, in order to include more code examples in the book.
No detailed applied discussions: each chapter contains only a simple interpretation of the results.
The first two chapters should be read (at least superficially) before any of the other chapters, but otherwise, the chapters can be read in any order.
The book is centered around
R code, with the goal of being “useful”
to applied users:
All the code should run quickly, even for complex multi-likelihood space-time problems.
Additional reading is required to understand the statistical properties of the models, but the reader can use and expand on the code to study these statistical properties.
The codes in the different chapters are independent, allowing the user to copy the code from only one chapter.
This is advanced applied statistical modeling, hence some of the models can be difficult to work with. We will continue to develop this book based on user feedback, through the online repository http://www.r-inla.org/spde-book.