survey (version 3.1)

svysmooth: Scatterplot smoothing and density estimation

Description

Scatterplot smoothing and density estimation for probability-weighted data.

Usage

svysmooth(formula, design, method = c("locpoly","quantreg"), bandwidth,quantile,df, ...)
## S3 method for class 'svysmooth':
plot(x, which=NULL, type="l", xlabs=NULL, ylab=NULL,...)
## S3 method for class 'svysmooth':
lines(x,which=NULL,...)

Arguments

formula

One-sided formula for density estimation, two-sided for smoothing

design

Survey design object

method

local polynomial smoothing for the mean or regression splines for quantiles

bandwidth

Smoothing bandwidth for "locpoly"

quantile

quantile to be estimated for "quantreg"

df

Degrees of freedom for "quantreg"

which

Which plots to show (default is all)

type

as for plot

xlabs

Optional vector of x-axis labels

ylab

Optional y-axis label

...

More arguments

x

Object of class svysmooth

Value

• An object of class svysmooth, a list of lists, each with x and y components.

Details

svysmooth does one-dimensional smoothing. If formula has multiple predictor variables a separate one-dimensional smooth is performed for each one. For method="locpoly" the extra arguments are passed to locpoly from the KernSmooth package, for method="quantreg" they are passed to rq from the quantreg package.

See Also

svyhist for histograms

Examples

data(api)
 dstrat<-svydesign(id=~1,strata=~stype, weights=~pw, data=apistrat, fpc=~fpc)

 smth<-svysmooth(api00~api99+ell,dstrat, bandwidth=c(40,10))
 dens<-svysmooth(~api99, dstrat,bandwidth=30)
 qsmth<-svysmooth(api00~ell,dstrat, quantile=0.75, df=3,method="quantreg")

 plot(smth)
 plot(smth, which="ell",lty=2,ylim=c(500,900))
 lines(qsmth, col="red")

 svyhist(~api99,design=dstrat)
 lines(dens,col="purple",lwd=3)