sm.options: Set or return options of sm library

In sm: Smoothing Methods for Nonparametric Regression and Density Estimation

Set or return options of sm library

Description

This function provides a means to control the behaviour of the sm library such as the colour of the plotted lines, the size of the grid in 2-d estimation, the set of evaluations points, and many others. A list may be given as the only argument, or any number of arguments may be in the name=value form. If no arguments are specified then the function returns the current settings of all the arguments.

Usage

sm.options(...)

Arguments

...	A list may be given as the only argument, or any number of arguments may be in the name=value form. The valid names of the arguments are given below.

add

a logical value which controls whether the estimate is added to the current plot. Its default value is FALSE, which creates a new plot. This argument applies only with one-dimensional data or to contour and rgl created from two-dimensional data.

alpha

a value, or a vector of values, lying between 0 and 1, which controls the transparency of the surfaces used to construct an rgl plot, when this form if display is requested. In the case of regression with two covariates, a single value for alpha is used and the default is 0.7. In the case of density estimation with three variables, alpha should be set to a vector whose length matches the number of contours to be drawn (specified by the props argument). In this case the default is seq(1, 0.5, length = length(props)).

alpha.mesh

a parameter, lying between 0 and 1, which controls the transparency of the mesh lines used to construct an rgl plot for regression with two covariates. The default value is 1.

band

a logical value which controls whether the distance between the nonparametric estimate and a reference model should be indicated as a band (one covariate), or through colour painting of the regression surface (two covariates). This is activated only when a model has been nominated through the model parameter. In the case of two covariates, the setting of the argument col has priority over band. The setting se = TRUE can also activate this feature.

col

the colour used for plotting observed points and estimated curves. Where groups are used, with one covariate, col may be set to a vector of colours associated with the groups. In regression with two covariates using an rgl display, col may be set to a single colour, or to the values "height" or "se". These latter two setting cause the surface to be painted according to its height or to standard error information; see the information on the parameters se, se.breaks and model. In the case of density estimation with three variables, col can be set to a vector whose length matches the number of contours to be drawn (specified by the props argument).

col.band

the colour used for the reference band when a model is specified in regression with one covariate or in density estimation with a single variable. Default: col.band="cyan".

col.mesh

the colour used for the ‘wire mesh’ representation plotting observed points in an rgl display for regression with two covariates. This can also be set by the second component of a vector of length two which is set for col. Default: col.mesh="black".

col.palette

the colours used for shading an image plot, or for surface painting in an rgl display, for regression with two covariates. Default: col.palette=topo.colors(12).

col.points

the colour used for plotting observed points in a regression with one covariate or an rgl display for regression with two covariates. Default: col.points="black".

delta

in sm.density, a value which will be added to the data before they are log transformed in the procedure to handle positive data. The value of delta is used only when positive takes the value TRUE. The default value is the smallest value observed in each dimension. This argument does not apply with three-dimensional data. Default: delta=NA

describe

logical flag which affects the behaviour of sm.script and provide.data. If describe=TRUE (default), a data documentation file is printed.

df

approximate degrees-of-freedom of the smoothing parameter used in sm.regression, when a numerical value of h is not specified. In this case, the equivalent value of h will be computed and included in the list returned on exit from sm.regression. Default value is 6 if x is a vector and 12 if x is a matrix.

diff.ord

in sm.regression, an integer defining the degree of differencing to be applied in the estimation process. When this argument is set to 1, the method of Rice, based on the squared differences of pairs of neighbouring observations, is used. When the argument is set to 2 (default), the method of Gasser, Sroka and Jennen-Steinmetz, based on differences between each observation and a linear interpolation from its two neighbours, is used.

display

This argument applies only with one- or two-dimensional data. The setting "none" will prevent any graphical output from being produced. In one dimensions, the default setting "line" will produce the estimate. (For compatibility with earlier versions of the package, the setting "se" will produce a variability band to show the variability, but not the bias, of the estimate. This should now be controlled by setting the separate parameter se to TRUE.) In two dimensions, the default setting "persp" will produce a perspective plot of the estimate, while the settings "slice", "image" and "rgl" will produce slice (contour), image or rgl plots.

eval.grid

logical flag which controls how the options eval.points are used for two-dimensional data. If eval.grid=TRUE (default), evaluation is performed at points obtained by the cross-product of the two columns of eval.points. If eval.grid=FALSE then evaluation is performed at points with coordinates specified by the coordinates in eval.points.

eval.points

the points at which the density or the regression curve or surface estimate should be evaluated, for the values returned in the result of the function. This should be a vector for one-dimensional data and a two-column matrix for two-dimensional data. This argument does not apply with three-dimensional data.

h.weights

a vector of weights which multiply the smoothing parameter used in the kernel function at each observation. This argument does not apply with three-dimensional data. Default value: 1.

hmult

a factor which can be used to multiply the normal smoothing parameter before construction of the density estimate. Default value: 1.

hull

a logical value which controls whether the estimate is evaluated and plotted only on grid points which fall within the convex hull of the data. When this argument is set to FALSE, evaluation and plotting take place at all grid points where the contribution from at least one kernel function is non-negligible. Both of these settings ensure that the estimate is not evaluated at points where there are no observations nearby. This argument applies only to sm.regression and sm.discontinuity in the case of two covariates.

lty

the line type used to plot the estimate. This argument applies only when the estimate is displayed as a curve or a contour. Default value: 1.

method

the method used to select smoothing parameters. In density estimation the default is "normal" which uses a value which is asymptotically optimal for the normal distribution. Other possibilities are "cv" for cross-validation and, for one-dimensional data only, "sj" for the Sheather-Jones method.

In nonparametric regression, the default is "df" which selects a smoothing parameters associated with the approximate degrees of freedom given in the df option. Other possibilities are "cv" for cross-validation and "aicc" for an AIC-based method proposed by Hurvich, Simonoff and Tsai.

nbins

the number of bins used in one-dimensional binning operations; in two-dimensional cases, nbins refers to the number of bins formed along each axis. Bins with 0 observed frequencies are ignored. If nbins=0, binning is not performed; if nbins=NA (default), binning is switched on when the number of observations exceeds a certain threshold, which depends on the function.

nboot

number of samples generated in bootstraps. Default value: 100.

ngrid

the number of points in the regular grid used to plot the estimate. For two- and three-dimensional data, ngrid refers to the number of points along the axis in each dimension. The same parameter is also used by a few other functions which perform some form of search (e.g. hcv). Default value for sm.regression: 50 and 20 for 1-, 2-dimensional data, respectively. Default value for sm.density: 100, 50 and 20 for 1-, 2- and 3-dimensional data, respectively.

panel

a logical value which, when set to true, creates a panel which allows interactive control of sm.regression or sm.density plots for one- or two-dimensional data. The panel can be used to alter the value of the smoothing parameter and control a variety of other settings.

panel.plot

a logical value which, when set to true (the default), places the plot inside the control panel (see the panel argument above), This creates a neater screen arrangement.

pch

the standard plotting character identified for data plotting. Default value: 1.

period

a vector of length one or two identifying the period for covariates which are on a periodic scale. Periodic smoothing is implemented by local mean estimation, using a von Mises kernel function. Non-periodic covariates are identified by NA. Default value: NA.

phi

the vertical rotation (in degrees) of perspective plots of estimate in the form of surfaces. Default value: 40.

poly.index

an integer defining local constant (0) or local linear (1) smoothing. Default value: 1.

positive

a logical value which indicates whether the data should be assumed to take positive values only, in sm.density. When this argument is set to TRUE, a log transformation is applied to the data before construction of a density estimate. The result is transformed back to the original scale. This argument does not apply with three-dimensional data. Default value: FALSE.

props

a vector defining the proportions of the data to be included within each contour in a slice plot, from two-dimensional data, or a contour surface plot, from three-dimensional data. In the three-dimensional case only the first element of the vector will be used. This argument does not apply to one-dimensional data. Default value: c(75,50,25).

rugplot

logical flag which regulates whether a rugplot is superimposed to the density estimate, in the univariate case. Default value: TRUE.

se

logical flag which regulates whether a standard error information is added to the plot produced by sm.regression. If a model is specified, then these standard errors refer to the difference between this fitted model and the nonparametric regression estimate. Default value: TRUE.

se.breaks

a numerical vector which defines the cut-points, on a standard error scale, for the assignment of colours when painting a regression surface with standard error information. Default value: c(-3, -2, 3, 3).

show.script

logical flag which affects the behaviour of sm.script when this is called with non-empty argument. If show.script=TRUE (default) a window is opened to display the source code of the script.

size

an integer which defines the size of plotted points in rgl displays. The default value is 2.

structure.2d

the structure of the smoothing parameter in two-dimensional settings. The default is "scaled", which uses the structure (h*sd(x[,1]), h*sd(x[,2])). Other possibilities are "separate", which uses (h1, h2), and "common" which uses (h, h). The "common" option may be particularly appropriate when the data have a spatial origin, where distances in each variable have the same meaning. Note that the "separate" option is not available when "method" is set to "df".

test

a logical flag controlling the production of a formal test, using the reference model as the null hypothesis. Default value: TRUE.

theta

the horizontal rotation (in degrees) of perspective plots of estimates in the form of surfaces. Default value: -30.

verbose

regulates the amount of messages and other output printed out. If verbose=0 only errors produce messages; if verbose=1 (default value) warnings and the more relevant numerical output are printed ; if verbose=2 more messages and more numerical output are printed.

xlab

the label attached to the x-axis.

xlim

the range of the horizontal axis of the plot. This argument does not apply with three-dimensional data.

yht

the upper limit of the vertical axis in a plot of a one-dimensional density estimate. The lower limit is always set to 0. This argument does not apply with two- or three-dimensional data.

ylab

the label attached to the y-axis.

ylim

the range of the vertical axis of the plot. This argument does not apply with three-dimensional data.

zlab

the label attached to the z-axis (three-dimensional plots only).

zlim

the range of the vertical axis when estimates are displayed as perspective plots.

Details

Arguments which are set by a function call will remain in effect until the end of the current S-plus session, unless overwritten by a subsequent call. In addition, they can be added as optional parameters of calls to specific functions of the sm package; in this case, their effect is limited to that function call.

See the documentation of specific functions for the list of options which are recognised by that function. Notice that some options are relevant only to some functions.

Value

a list with the updated values of the parameters. If the argument list is not empty, the returned list is invisible.

Examples

## Not run: sm.options(poly.index = 0)
# subsequent regression estimations will be performed using local means
# instead of local regression
#
sm.options(describe = FALSE)  
# turns off typing documentation files of data loaded by `sm.script'
# (works from command-line)
# 

## End(Not run)

sm documentation built on May 29, 2024, 2:28 a.m.