Type: Package
Title: R Utilities Accompanying the Software Package BayesX
Version: 0.3-3
Date: 2023-10-20
Description: Functions for exploring and visualising estimation results obtained with BayesX, a free software for estimating structured additive regression models (https://www.uni-goettingen.de/de/bayesx/550513.html). In addition, functions that allow to read, write and manipulate map objects that are required in spatial analyses performed with BayesX.
Depends: shapefiles
Imports: sp, sf, colorspace, coda, splines, methods, interp
Suggests: spdep
LazyLoad: yes
License: GPL-2 | GPL-3
NeedsCompilation: no
Packaged: 2023-10-20 06:19:08 UTC; nik
Author: Nikolaus Umlauf ORCID iD [aut, cre], Thomas Kneib [aut], Nadja Klein [aut], Felix Heinzl [ctb], Andreas Brezger [ctb], Daniel Sabanes Bove [ctb]
Maintainer: Nikolaus Umlauf <Nikolaus.Umlauf@uibk.ac.at>
Repository: CRAN
Date/Publication: 2023-10-20 07:20:07 UTC

R Utilities Accompanying the Software Package BayesX

Description

This package provides functionality for exploring and visualising estimation results obtained with the software package BayesX for structured additive regression. It also provides functions that allow to read, write and manipulate map objects that are required in spatial analyses performed with BayesX.

Author(s)

Nikolaus Umlauf, Thomas Kneib, Nadja Klein, Felix Heinzl, Andreas Brezger, Daniel Sabanes Bove

References

Belitz C, Brezger A, Kneib T, Lang S (2011). BayesX - Software for Bayesian Inference in Structured Additive Regression Models. Version 2.0.1. URL https://www.uni-goettingen.de/de/bayesx/550513.html.

Compute direction of points in a polygon.

Description

This is (private, not exported) R code replacing the call to maptools:::.ringDirxy, which links to C code.

Usage

.ringDirxy(xy)

Arguments

xy

a two-column matrix with x and y coordinates of at least 3 points which form a polygon

Value

Returns the integer -1 if the direction is counter-clockwise and 1 if the direction is clockwise.

Note

The function uses vectorization and should not be very much slower than the original.

Author(s)

Daniel Sabanes Bove

Examples

## Not run: ## Create test polygon ordered clockwise.
nPoints <- 200
sequence <- seq(1, 10, length = nPoints)
firstHalf <- sequence[1:(nPoints/2)]

test <- cbind(
  "x" = c(sequence, rev(sequence)),
  "y" = c(firstHalf, rev(firstHalf), -firstHalf, -rev(firstHalf))
)

plot(test, type="n")
polygon(test)

## Compute direction with this and the original function,
BayesX:::.ringDirxy(test)

## End(Not run)

Add Neighborhood Relations

Description

Adds a neighborhhod relationship between two given regions to a map object in graph format.

Usage

add.neighbor(map, region1, region2)

Arguments

map

Map object in graph format that should be modified.

region1, region2

Names of the regions that should be connected as neighbors.

Value

Returns an adjacency matrix that represents the neighborhood structure of map plus the new neighborhood relation in graph format.

Author(s)

Felix Heinzl, Thomas Kneib

See Also

get.neighbor,delete.neighbor,read.gra,write.grabnd2gra

Examples

germany <- read.gra(system.file("examples/germany.gra", package="BayesX"))
get.neighbor(germany, c("1001", "7339"))
germany <- add.neighbor(germany, "7339", "1001")
get.neighbor(germany, c("1001", "7339"))

Convert Boundary Format to Graph Format

Description

Converts a map in boundary format to a map in graph format.

Usage

bnd2gra(map)

Arguments

map

Map in boundary format that should be converted.

Value

Returns an adjacency matrix that represents the neighborhood structure of the map object in graph format.

Author(s)

Felix Heinzl, Thomas Kneib

References

BayesX Reference Manual. Available from https://www.uni-goettingen.de/de/bayesx/550513.html

See Also

read.bnd,read.gra,write.bnd,write.gra

Examples

tanzania.bnd <- read.bnd(system.file("examples/tanzania.bnd", package="BayesX"))
tanzania.gra <- bnd2gra(tanzania.bnd)

Create map objects for some points and a given distance

Description

Creates a map object from a list of coordinates by treating observations within a certain distance as neighbors. The resulting neighborhood structure is stored in a map object in graph format while a map in boundary format is created to enable visualisation.

Usage

createxymap(x, y, districts=NULL, p=2, max.dist)

Arguments

x

Vector of x-coordinates of underlying points

y

Vector of y-coordinates of underlying points

districts

Either NULL or a vector of names for labeling points. If districts=NULL, points are labelled by index.

p

Any p-norm with p>=1 can be chosen as the distance measure with the Euclidian distance (p=2) being the default. For p=Inf, the maximum of coordinates is used. Except for p=Inf, calculations can fail for huge p.

max.dist

Value which determines the neighborship. Points whose distance is smaller or equal than max.dist are considered as neighbors.

Value

List of two elements: map object in graph format and map object in boundary format.

Author(s)

Felix Heinzl, Thomas Kneib

See Also

read.gra,read.bnd,drawmap

Examples

x <- c(3,3,5,9.5,11,11)
y <- c(2,4,4,6,4.5,5)
xymap <- createxymap(x, y, districts=c("A","B","C","D","E","F"), max.dist=2)
xymap$gra
drawmap(map=xymap$bnd)

Delete Neighborhood Relations

Description

Adds the neighborhhod relationship between two given regions from a map object in graph format.

Usage

delete.neighbor(map, region1, region2)

Arguments

map

Map object in graph format that should be modified.

region1, region2

Names of the regions that should no longer be regarded as neighbors.

Value

Returns an adjacency matrix that represents the neighborhood structure of map minus the deleted neighborhood relation in graph format.

Author(s)

Felix Heinzl, Thomas Kneib

See Also

get.neighbor,add.neighbor,read.gra,write.grabnd2gra

Examples

germany <- read.gra(system.file("examples/germany.gra", package="BayesX"))
get.neighbor(germany, c("7339"))
germany <- delete.neighbor(germany, "7339", "7141")
get.neighbor(germany, c("7339"))

Drawing Geographical Information

Description

Visualises variables that are spatially aligned according to a given map object. Each of the regions in a map will be coloured accoring to the value of the variable.

Usage

drawmap(data, map, regionvar=2, plotvar=3, limits, cols="hcl", nrcolors=100, 
        swapcolors=FALSE, pcat=FALSE,
        hcl.par=list(h=c(120, 0), c=60, l=c(45,60), power=1.2), 
        hsv.par=list(s=1, v=1), legend=TRUE, drawnames=FALSE, cex.names=0.7, 
        cex.legend=0.7, mar.min=2, density=15, ...)

Arguments

data

Either the name of a file or a data frame containing the variables to be visualised. If missing, the map will be visualised without superposition of any further information

map

Map object containing the required boundary information (as obtained by a call to read.bnd

regionvar

Defines the variable specifying the geographical regions. Either the name of a variable in data or the index of the corresponding column.

plotvar

Defines the variable that should be visualised. Either the name of a variable in data or the index of the corresponding column.

limits

Restricts (or extends) the coloring scheme to a range of values.

cols

Color scheme to be employed. Could be either a vector of colors or one out of the following pre-defined schemes: hcl, hsv, grey

nrcolors

Number of colors (only meaningful when using one of the pre-defined colour schemes).

swapcolors

Reverse the order of colors (works also with user-specified colours but will be most usefule with the pre-defined schemes).

pcat

Option for the visualisation of posterior probabilities. In this case, a three-colour scheme representing significantly positive, insignificant and significantly negative values.

hcl.par

Parameters for the hcl colour scheme (see the documentation of diverge_hcl in package vcd for details).

hsv.par

Parameters for the hsv colour scheme (see the documentation of hsv for details).

legend

Should a legend be added to the figure?

drawnames

Adds the name of each region as a text label to the plot. In most cases the result will be confusing but may be useful when checking the validity of a map.

cex.names

Magnification to be used for the names (if drawnames=TRUE).

cex.legend

Magnification to be used for the legend.

mar.min

Controls the definition of boundaries. Could be either NULL for individual settings of mar or a value which defines mar as follows: The boundaries will be calculated according to the height to width ratio of the map with minimal boundary mar.min.

density

Regions without data will be visualised with diagonal stripes. density defines how dense the stripes should be.

...

Further arguments to be passed to the plot calls that visualise the region boundaries (probably not useful at all).

Author(s)

Felix Heinzl, Thomas Kneib, Andreas Brezger

See Also

read.bnd

Examples

germany <- read.bnd(system.file("examples/germany.bnd", package="BayesX"))
drawmap(map=germany)
drawmap(map=germany, drawnames=TRUE)

res <- read.table(system.file("examples/spatial_f_regions_spatial.res", 
                              package="BayesX"), header=TRUE)

drawmap(res, map=germany)
drawmap(res, map=germany, limits=c(-2,4))
drawmap(res, map=germany, regionvar="regions", plotvar="pmed")
drawmap(res, map=germany, legend=FALSE)
drawmap(res, map=germany, legend=FALSE, main="spatial effect")

drawmap(res, map=germany, cols="hsv")
drawmap(res, map=germany, swapcolors=TRUE, cols="hsv")
drawmap(res, map=germany, cols="grey")
drawmap(res, map=germany,
        cols=c('darkgreen','green','yellow','orange','red','darkred'))

drawmap(res, map=germany, pcat=TRUE, cols="hcl")
drawmap(res, map=germany, pcat=TRUE, cols="hsv")
drawmap(res, map=germany, pcat=TRUE, cols="grey")

drawmap(res, map=germany, nrcolors=10, cols="hcl")
drawmap(res, map=germany, nrcolors=10, cols="hsv")
drawmap(res, map=germany, nrcolors=10, cols="grey")

drawmap(res, map=germany, cols="hcl",
        hcl.par=list(h=c(0,120), c=60, l=c(45,90), power=1.2))
drawmap(res, map=germany, cols="hcl",
        hcl.par=list(h=c(300,120), c=60, l=c(45,90), power=1.2))
drawmap(res, map=germany, cols="hcl",
        hcl.par=list(h=c(40,260), c=60, l=c(45,90), power=1.2))
drawmap(res, map=germany, cols="hsv", hsv.par=list(s=0.7, v=0.7))

Extract MCMC samples from a BayesX results directory

Description

This is a convenience function to extract samples from a BayesX results directory, which processes the log file to e.g. convert the spline coefficients samples to function values samples.

Usage

extractSamples(directoryWithBasename, 
               logfile = file.path(dirname(directoryWithBasename), "log.txt"))

Arguments

directoryWithBasename

The BayesX results directory with basename for the files (e.g. "results/test", if this was specified as outfile in BayesX for the bayesreg object)

logfile

The log file of the MCMC run, defaults to log.txt in the results directory.

Value

Returns a list with the extracted samples of effects and deviances as well as the prediction data.frame:

<function name>

for P-Splines, Random Walks and spatial effects: a list with mcmc objects 'functionSamples' and 'varianceSamples' containing the respective effects/function and variance parameter samples.

FixedEffects

an mcmc object of all fixed simple parametric effects

RandomEffects

if there is at least one random effect in the model, this is a list, with elements in the first hierarchy being the group ID names, and elements in the second hierarchy being the names of the covariates. The leafs are the mcmc objects 'functionSamples' and 'varianceSamples', as for the other non-fixed terms

Deviance

an mcmc object with the (unstandardized and saturated) deviance

means

if the option predictmu was used, this mcmc object contains the mean samples

scale

an mcmc object with the possible scale parameter samples

lassoCoefficients

an mcmc object with the possible lasso regression parameter samples

ridgeCoefficients

an mcmc object with the possible ridge regression parameter samples

PredictMeans

data.frame corresponding to the possible predictmean file in the BayesX directory

Additionally, entries for possibly remaining lasso or ridge variance parameters etc. are included in the return list.

Warning

You should be sure that only one MCMC run is saved in the given results directory in order to get sensible results out of this function.

Author(s)

Daniel Sabanes Bove, with contributions by Fabian Scheipl

Examples

## get the samples
samples <- extractSamples(file.path(system.file("examples/samples", package="BayesX"),
                                     "res"))
str(samples)

## check deviance convergence
plot(samples$Deviance)

## fixed parametric effects
plot(samples$FixedEffects)

## nonparametric effects:

## handy plot function to get means and pointwise credible intervals
nonpPlot <- function(samplesMatrix,
                     ...)
{
    x <- as.numeric(colnames(samplesMatrix))

    yMeans <- colMeans(samplesMatrix)
    yCredible <- t(apply(samplesMatrix,
                         MARGIN=2,
                         FUN=quantile,
                         prob=c(0.025, 0.975),
                         na.rm=TRUE))
    
    matplot(x, cbind(yMeans, yCredible),
            type="l",
            lty=c(1, 2, 2),
            lwd=c(2, 1, 1),
            col=c(1, 2, 2),
            ...)
}

nonpPlot(samples$f_x1$functionSamples,
         xlab=expression(x[1]),
         ylab=expression(hat(f)(x[1])))
nonpPlot(samples$f_x2$functionSamples,
         xlab=expression(x[2]),
         ylab=expression(hat(f)(x[2])))

## spatial effect
tanzania <- read.bnd(file=system.file("examples/tanzania.bnd", package="BayesX"))
drawmap(map=tanzania,
        data=
        with(samples$f_district,
             data.frame(name=colnames(functionSamples),
                        estimate=colMeans(functionSamples))),
        regionvar="name",
        plotvar="estimate")

Combine Regions

Description

Combines a list of several regions of a map object in boundary format into a single region.

Usage

fuse(map, regions, name)

Arguments

map

Map object in boundary format that should be modified.

regions

Vector of regions to be combined

name

Name that should be given to the region arising from fusing the specified regions.

Value

Map object in boundary format with the specified regions combined.

Author(s)

Nadja Klein

See Also

read.bnd,write.bnd

Examples

## Not run: map <- read.bnd(system.file("examples/germany9301.bnd",
  package = "BayesX"))
drawmap(map = map, drawnames = TRUE)

## Vector of regions to be combined.
regions <- c("1056","1060","1061")

## New name of combined region.
newname <- "1"
newmap <- fuse(map,regions,newname)
drawmap(map = newmap, drawnames = TRUE)

## Vector of regions to be combined.
germany <- read.bnd(system.file("examples/germany.bnd", package="BayesX"))
drawmap(map = germany, drawnames = TRUE)
regions <- c("9371","9373","9374","9471","9472","9474","9574")

## New name of combined region.
newname <- "1"
newmap <- fuse(germany, regions, newname)
drawmap(map = newmap, drawnames = TRUE)

## End(Not run)

Compute Centroids of Polygons

Description

Computes all areas and centroids of the regions of a given map in boundary format.

Usage

get.centroids(map)

Arguments

map

Map object in boundary format.

Value

Matrix of area and centroids.

Author(s)

Felix Heinzl, Thomas Kneib

Examples

germany <- read.bnd(system.file("examples/germany.bnd", package="BayesX"))
centroids <- get.centroids(germany)
centroids[1:10,]

plot(c(2100,3700),c(6800,8500),type="n", xlab="", ylab="")
for(i in 1:10){
   polygon(germany[[i]])
   region <- attr(germany,"names")[i]
   text(x=centroids[i,2]+50, y=centroids[i,3]+30, region, cex=0.7)
}   
points(centroids[1:10,2:3], col='red', pch=16)

Obtain Neighbors of Given Regions

Description

Extracts the neighbors of a number of regions from a map in graph format.

Usage

get.neighbor(map, regions)

Arguments

map

Map object in graph format.

regions

Vector of names of regions for which the neighbors should be axtracted.

Value

A list of vectors containing the neighbors of the elements in regions.

Author(s)

Felix Heinzl, Thomas Kneib

See Also

add.neighbor,delete.neighbor

Examples

germany <- read.gra(system.file("examples/germany.gra", package="BayesX"))
get.neighbor(germany, "1001")
get.neighbor(germany, c("1001", "7339"))

Extract numbers from (BayesX log file) strings

Description

This is an internal helper function for extractSamples which extracts numbers from (BayesX log file) strings.

Usage

getNumbers(beforeStringsList, stringVector)

Arguments

beforeStringsList

The list with the strings standing before the numbers which shall be extracted.

stringVector

The vector of strings to be searched for the numbers.

Value

Returns a list with the extracted numbers.

Author(s)

Daniel Sabanes Bove

Examples

## create a nice example
sampleStrings <- c("Second: 385",
                   "  First:  70000 ",                   
                   "asdfkf T: 24      ")

## test the function
extractedNumbers <- BayesX:::getNumbers(beforeStringsList=
                                        list(first="First:",
                                             second="Second:",
                                             third="T:"),
                                        stringVector=sampleStrings)

## and assert that the result is as expected
stopifnot(identical(extractedNumbers,
                    list(first=70000, second=385, third=24)))

## specific test
stopifnot(identical(BayesX:::getNumbers(beforeStringsList=
                                        list(Iterations = "Number of iterations:"),
                                        stringVector=
                                        "  Number of iterations: 70000 "),
                    list(Iterations=70000)))

Computing Highest Posterior Density (HPD) Intervals

Description

Compute approximate HPD intervals out of MCMC-samples in BayesX

Usage

hpd(data, alpha = 0.05, ...)
hpd.coda(data, alpha = 0.05)

Arguments

data

Either the name of a file or a data frame containing the sample.

alpha

A numeric scalar in the interval (0,1) such that 1 - alpha is the target probability content of the intervals.. The default is alpha = 0.05.

...

Further parameters to be passed to the internal call of optim and integrate.

Details

hpd computes the HPD interval based on a kernel density estimate of the samples. hpd.coda computes the HPD interval with the function HPDinterval available in package coda.

Author(s)

Nadja Klein

Examples

res <- read.table(system.file("examples/nonparametric_f_x_pspline_sample.raw",
  package="BayesX"), header = TRUE)
hpd(res)
hpd.coda(res)

Convert nb and gra format into each other

Description

Convert neighborhood structure objects of class "nb" from R-package spdep to graph objects of class "gra" from R-package BayesX and vice versa.

Usage

nb2gra(nbObject)
gra2nb(graObject)

Arguments

nbObject

neighborhood structure object of class "nb"

graObject

graph object of class "gra"

Value

Equivalent object in the other format.

Author(s)

Daniel Sabanes Bove

See Also

sp2bnd, bnd2sp for conversion between the geographical information formats and read.gra, write.gra for the interface to the BayesX files.

Examples

## Not run: ## first nb to gra:
if(requireNamespace("sf") &
   requireNamespace("spdep")) {
  library("sf")
  library("spdep")

  columbus <- st_read(system.file("etc/shapes/columbus.shp",
    package = "spdep")[1])
  colNb <- poly2nb(columbus)
  ## ... here manual editing is possible ...
  ## then export to graph format
  colGra <- nb2gra(colNb)

  ## and save in BayesX file
  graFile <- tempfile()
  write.gra(colGra, file=graFile)

  ## now back from gra to nb:
  colGra <- read.gra(graFile)
  newColNb <- gra2nb(colGra)
  newColNb
  ## compare this with the original
  colNb
  ## only the call attribute does not match (which is OK):
  all.equal(newColNb, colNb,
    check.attributes=FALSE)
  attr(newColNb, "call")
  attr(colNb, "call")
}

## End(Not run)

Convert sp and bnd format into each other

Description

Convert geographical information objects of class "SpatialPolygons" (or specializations) from R-package sp to objects of class "bnd" from R-package BayesX and vice versa.

Usage

sp2bnd(spObject, regionNames, height2width, epsilon)
bnd2sp(bndObject)

Arguments

spObject

object of class "SpatialPolygons" (or specializations)

regionNames

character vector of region names (parallel to the Polygons list in spObject), defaults to the IDs

height2width

ratio of total height to width, defaults to the bounding box values

epsilon

how much can two polygons differ (in maximum squared Euclidean distance) and still match each other?, defaults to machine precision

bndObject

object of class "bnd"

Value

Equivalent object in the other format.

Author(s)

Daniel Sabanes Bove

See Also

nb2gra, gra2nb for conversion between the neighborhood structure formats and read.bnd, write.bnd for the interface to the BayesX files.

Examples

## Not run: ## bnd to sp:
germany <- read.bnd(system.file("examples/germany2001.bnd", package="BayesX"))
spGermany <- bnd2sp(germany)

## plot the result together with the neighborhood graph
library(sp)
plot(spGermany)
library(spdep)
nbGermany <- poly2nb(spGermany)
plot(nbGermany, coords=coordinates(spGermany), add=TRUE)

## example with one region inside another
spExample <- spGermany[c("7211", "7235"), ]
plot(spExample)
plot(poly2nb(spExample), coords=coordinates(spExample), add=TRUE)

## now back from sp to bnd:
bndGermany <- sp2bnd(spGermany)
drawmap(map=bndGermany)

## compare names and number of polygons
stopifnot(identical(names(bndGermany),
                    names(germany)),
          identical(length(bndGermany),
                    length(germany)))

## compare contains-relations
surrounding <- attr(bndGermany, "surrounding")
whichInner <- which(sapply(surrounding, length) > 0L)
bndContainsData <- data.frame(inner=names(bndGermany)[whichInner],
                              outer=unlist(surrounding))

surrounding <- attr(germany, "surrounding")
whichInner <- which(sapply(surrounding, length) > 0L)
originalContainsData <- data.frame(inner=names(germany)[whichInner],
                                   outer=unlist(surrounding))

stopifnot(all(bndContainsData[order(bndContainsData$inner), ] ==
              originalContainsData[order(originalContainsData$inner), ]))

## End(Not run)

Computing and Plotting Autocorrelation Functions

Description

Computes and plot autocorrelation functions for samples obtained with MCMC in BayesX

Usage

plotautocor(data, ask = TRUE, lag.max=100, ...)

Arguments

data

Either the name of a file or a data frame containing the sample.

ask

plotautocor will plot separate autocorrelation functions for each parameter. If ask=TRUE, the user will be prompted before showing the next plot.

lag.max

Maximum number of lags to be considered.

...

Further parameters to be passed to the internal call of plot such as ylim, etc.

Author(s)

Felix Heinzl, Thomas Kneib

Examples

res <- read.table(system.file("examples/nonparametric_f_x_pspline_sample.raw", 
                              package="BayesX"), header=TRUE)
plotautocor(res)
plotautocor(res, lag.max=50)

Plotting Nonparametric Function Estimates

Description

Plots nonparametric function estimates obtained from BayesX

Usage

plotnonp(data, x = 2, y = c(3, 4, 5, 7, 8), ylim = NULL, 
         lty = c(1, 2, 3, 2, 3), cols = rep(1, length(y)), month, year, step=12, 
         xlab, ylab, ...)

Arguments

data

Either the name of a file or a data frame containing the estimation results.

x

Defines the x-axis in the plot. Either the name of a variable in data or the index of the corresponding column.

y

Defines the variables to be plotted against x. May be either a vector of names of variables in data or the corresponding indices. The default choice corresponds to the point estimate plus two confidence bands.

ylim

Since plotnonp plots multiple y-variables, it automatically determines the appropriate ylim to make all curves visible. Argument ylim allows to override this default behaviour with fixed values.

lty

Vector of line types used for plotting (must have the same length as y). The default corresponds to solid lines for the point estimate and dashed and dotted lines for the confidence bands.

cols

Vector of colors used for plotting (must have the same length as y). Default are black lines.

month, year, step

Provide specific annotation for plotting estimation results for temporal variables. month and year define the minimum time point whereas step specifies the type of temporal data with step=4, step=2 and step=1 corresponding to quartely, half yearly and yearly data.

xlab, ylab

plotnonp constructs default labels that can be overwritten by these arguments

...

Further arguments to be passed to the interval call of plot such as type, etc.

Author(s)

Felix Heinzl, Andreas Brezger and Thomas Kneib

See Also

drawmap,plotautocor,plotsample,plotsurf

Examples


res <- read.table(system.file("examples/nonparametric_f_x_pspline.res", 
                              package="BayesX"), header=TRUE)
plotnonp(res)
plotnonp(res, x="x")
plotnonp(res, x="x", y="pmean")
plotnonp(res, x="x", y="pmed")
plotnonp(res, x="x", y="pmed", ylim=c(-2,2))
plotnonp(res, x="x", y=c("pmean", "pqu10", "pqu90"), lty=c(1,1,1), 
         col=c("red","blue","blue"))
plotnonp(res, xlab="some variable", ylab="f(some variable)", 
         main="Nonlinear effect of some variable", sub="penalised spline")

res <- read.table(system.file("examples/nonparametric2_f_time_pspline.res", 
                              package="BayesX"), header=TRUE)
plotnonp(res)
plotnonp(res, month=1, year=1980, step=12)

res <- res[1:18,]                                           
plotnonp(res, month=1, year=1980, step=12)

Plotting Sampling Paths

Description

Plots sampling paths obtained with MCMC-sampling in BayesX

Usage

plotsample(data, ask = TRUE, ...)
plotsample.coda(data, ask = TRUE, ...)

Arguments

data

Either the name of a file or a data frame containing the sample.

ask

plotsample will plot separate sampling paths for each parameter. If ask=TRUE, the user will be prompted before showing the next plot.

...

Further parameters to be passed to the internal call of plot such as ylim, etc.

Details

plotsample simply plots sampling paths while plotsampe.coda makes use of the plotting facilities available in package coda.

Author(s)

Felix Heinzl, Andreas Brezger, Thomas Kneib

See Also

drawmap,plotautocor,plotnonp,plotsurf,

Examples

res <- read.table(system.file("examples/nonparametric_f_x_pspline_sample.raw",
                              package="BayesX"), header=TRUE)
plotsample(res)

Visualise Surface Estimates

Description

Visualises surface estimates obtained with BayesX.

Usage

plotsurf(data, x=2, y=3, z=4, mode=1, ticktype="detailed", 
         expand=0.75, d=100, theta=-30, phi=25, ...)

Arguments

data

Either the name of a file or a data frame containing the estimation results.

x

Defines the x-axis in the plot. Either the name of a variable in data or the index of the corresponding column.

y

Defines the y-axis in the plot. Either the name of a variable in data or the index of the corresponding column.

z

Defines the z-axis in the plot. Either the name of a variable in data or the index of the corresponding column.

mode

plotsurf is mostly a convenient interface to the functions persp (mode=1), image (mode=2) and contour (mode=3).

ticktype, expand, d, theta, phi

Overwrite the default behaviour of persp

...

Further parameteres that are parsed to the internal call to persp, image or contour

Author(s)

Felix Heinzl, Thomas Kneib

See Also

drawmap,plotautocor,plotsample,plotnonp

Examples

res <- read.table(system.file("examples/surface_f_x1_x2_pspline.res", 
                              package="BayesX"), header=TRUE)

plotsurf(res)
plotsurf(res, mode=2)
plotsurf(res, mode=3)

plotsurf(res, x="x1", y="x2", z="pmed")

plotsurf(res, ticktype="simple")

plotsurf(res, main="3D-Plot", xlab="myx", ylab="myy", zlab="f(myx,myy)")

Read Geographical Information in Boundary Format

Description

Reads the geographical information provided in a file in boundary format (see Ch. 5 of the BayesX Reference Manual) and stores it in a map object.

Usage

read.bnd(file, sorted=FALSE)

Arguments

file

Name of the boundary file to be read.

sorted

Should the regions be ordered by the numbers speciying the region names (sorted=TRUE)?

Value

Returns a list of polygons that form the map. Additional attributes are

surrounding

Parallel list where for each polygon, the name of a possible surrounding region is saved.

height2width

Ratio between height and width of the map. Allows customised drawing and storage in files by specifying the appropriate height and width.

class

Indicates whether the map is stored in boundary format (bnd) or graph format (gra). Maps returned by read.bnd are of class bnd

Author(s)

Daniel Sabanes Bove, Felix Heinzl, Thomas Kneib, Andreas Brezger

References

BayesX Reference Manual. Available from https://www.uni-goettingen.de/de/bayesx/550513.html

See Also

write.bnd,drawmap,read.gra,write.gra

Examples

germany <- read.bnd(system.file("examples/germany.bnd", package="BayesX"))
drawmap(map=germany)
attributes(germany)

germany <- read.bnd(system.file("examples/germany2001.bnd", package="BayesX"))
drawmap(map=germany)
attributes(germany)

Read Geographical Information in Graph Format

Description

Reads the geographical information provided in a file in graph format (see Ch. 5 of the BayesX Reference Manual) and stores it in a map object.

Usage

read.gra(file, sorted=FALSE)

Arguments

file

Name of the graph file to be read.

sorted

Should the regions be ordered by the numbers speciying the region names (sorted=TRUE)?

Value

Returns an adjacency matrix that represents the neighborhood structure defined in the graph file. Additional attributes are

dim

Dimension of the (square) adjacency matrix.

dimnames

List of region names corresponding to rows and columns of the adjacency matrix.

class

Indicates whether the map is stored in boundary format (bnd) or graph format (gra). Maps returned by read.gra are of class gra

Author(s)

Thomas Kneib, Felix Heinzl

References

BayesX Reference Manual. Available from https://www.uni-goettingen.de/de/bayesx/550513.html

See Also

write.gra,read.bnd,write.bnd,get.neighbor,add.neighbor,delete.neighbor

Examples

germany <- read.gra(system.file("examples/germany.gra", package="BayesX"))
attributes(germany)

convert a shape-file into a boundary object

Description

Converts the geographical information provided in a shape-file into a boundary object (see Ch. 5 of the Reference Manual)

Usage

shp2bnd(shpname, regionnames, check.is.in = TRUE)

Arguments

shpname

Base filename of the shape-file (including path)

regionnames

Either a vector of region names or the name of the variable in the dbf-file representing these names

check.is.in

Test whether some regions are surrounded by other regions (FALSE speeds up the execution time but may result in a corrupted bnd-file)

Value

Returns a boundary object, i.e. a list of polygons that form the map. See read.bnd for more information on the format.

Author(s)

Felix Heinzl, Daniel Sabanes Bove, Thomas Kneib with contributions by Michael Hoehle and Frank Sagerer

References

BayesX Reference Manual. Available from https://www.uni-goettingen.de/de/bayesx/550513.html

See Also

write.bnd,drawmap,read.bnd

Examples

## read shapefile into bnd object
shpName <- sub(pattern="(.*)\\.dbf", replacement="\\1",
               x=system.file("examples/northamerica_adm0.dbf",
                             package="BayesX")) 
north <- shp2bnd(shpname=shpName, regionnames="COUNTRY")

## draw the map
drawmap(map=north)

## compare with shipped bnd file
shippedBnd <- read.bnd(system.file("examples/northamerica.bnd", package="BayesX"))
stopifnot(all.equal(north, shippedBnd))

Round Boundary Information

Description

Rounds the boundary information in a map object in boundary format to a specified precision.

Usage

smooth.bnd(map, digits = 2, scale = 1)

Arguments

map

Map object in boundary format that should be modified.

digits

Number of digits to round to.

scale

Scaling factor that should be applied for rounding. For example, with scale=0.1 all polygons are divided by 10 before rounding.

Value

Map object in boundary format rounded to the specified precision.

Author(s)

Felix Heinzl, Thomas Kneib

See Also

read.bnd,write.bnd

Saving Maps in Boundary Format

Description

Writes the information of a map object to a file (in boundary format)

Usage

write.bnd(map, file, replace = FALSE)

Arguments

map

Map object ot be saved (should be in boundary format).

file

Name of the file to write to

replace

Should an existing file be overwritten with the new version?

Author(s)

Thomas Kneib, Felix Heinzl

References

BayesX Reference Manual. Available from https://www.uni-goettingen.de/de/bayesx/550513.html

See Also

write.gra,read.gra,read.bnd

Saving Maps in Graph Format

Description

Writes the information of a map object to a file (in graph format)

Usage

write.gra(map, file, replace = FALSE)

Arguments

map

Map object ot be saved (should be in graph format, see bnd2gra for the conversion of boundary format to graph format).

file

Name of the file to write to

replace

Should an existing file be overwritten with the new version?

Author(s)

Thomas Kneib, Felix Heinzl

References

BayesX Reference Manual. Available from https://www.uni-goettingen.de/de/bayesx/550513.html

See Also

write.bnd,read.gra,read.bnd