From 54c69dc5464d864aa0cdda65c627cdea703b10b6 Mon Sep 17 00:00:00 2001 From: Andrew Zammit Mangion Date: Wed, 27 Dec 2023 07:30:15 +1100 Subject: [PATCH] Fixed small mistake in FRK_intro.Rnw (missing escape characters) --- vignettes/FRK_intro.Rnw | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/vignettes/FRK_intro.Rnw b/vignettes/FRK_intro.Rnw index b971a78..39b55be 100644 --- a/vignettes/FRK_intro.Rnw +++ b/vignettes/FRK_intro.Rnw @@ -313,7 +313,7 @@ GridBAUs1 <- auto_BAUs(manifold = plane(), # 2D plane nonconvex_hull=FALSE) # convex hull @ -\noindent The \tt{auto\_BAUs} function takes several arguments (see \tt{help(auto\_BAUs)} for details). Above, we instruct the helper function to construct BAUs on the plane, centred around the data \tt{meuse} with each BAU of size 100 $\times$ 100 (with units in m since the data is supplied with x-y coordinates in m). The \tt{type="grid"} input instructs that we want a rectangular grid and not a hexagonal lattice (use \tt{"hex"} for a hexagonal lattice), and \tt{convex=-0.05} is a specific parameter controlling the buffer-width of the spatial-domain boundary. The name `convex' was chosen as it is also used to control the buffer in case a non-convex hull is desired by setting \tt{nonconvex\_hull=TRUE} (see \tt{fmesher::fm_nonconvex_hull_inla} for more details and note that {\bf fmesher} needs to be installed for this option to be set). For the $i$th BAU, we also need to attribute the element $\varv_i$ that describes the hetereoscedascity of the fine-scale variation for that BAU. As described in Section \ref{sec:SREModel}, this component encompasses all process variation that occurs at the BAU scale and only needs to be known up to a constant of proportionality, $\sigma^2_\xi$ or $\sigma^2_\delta$ (depending on the chosen model); this constant is estimated using maximum likelihood with \tt{SRE.fit} using the EM algorithm of Section \ref{sec:estimation}. Typically, geographic features such as altitude are appropriate, but in this case we will just set this parameter to unity. It is important that this field is labelled `fs': +\noindent The \tt{auto\_BAUs} function takes several arguments (see \tt{help(auto\_BAUs)} for details). Above, we instruct the helper function to construct BAUs on the plane, centred around the data \tt{meuse} with each BAU of size 100 $\times$ 100 (with units in m since the data is supplied with x-y coordinates in m). The \tt{type="grid"} input instructs that we want a rectangular grid and not a hexagonal lattice (use \tt{"hex"} for a hexagonal lattice), and \tt{convex=-0.05} is a specific parameter controlling the buffer-width of the spatial-domain boundary. The name `convex' was chosen as it is also used to control the buffer in case a non-convex hull is desired by setting \tt{nonconvex\_hull=TRUE} (see \tt{fmesher::fm\_nonconvex\_hull\_inla} for more details and note that {\bf fmesher} needs to be installed for this option to be set). For the $i$th BAU, we also need to attribute the element $\varv_i$ that describes the hetereoscedascity of the fine-scale variation for that BAU. As described in Section \ref{sec:SREModel}, this component encompasses all process variation that occurs at the BAU scale and only needs to be known up to a constant of proportionality, $\sigma^2_\xi$ or $\sigma^2_\delta$ (depending on the chosen model); this constant is estimated using maximum likelihood with \tt{SRE.fit} using the EM algorithm of Section \ref{sec:estimation}. Typically, geographic features such as altitude are appropriate, but in this case we will just set this parameter to unity. It is important that this field is labelled `fs': <<>>= GridBAUs1$fs <- 1 # fine-scale variation at BAU level