Package loct

September 20, 2011

Version 1.5-6 Title Local Regression, Likelihood and Density Estimation. Date 2010-01-20 Author Catherine Loader Maintainer Andy Liaw <[email protected]> Description Local regression, likelihood and density estimation. Depends R (>= 2.0.1), akima, lattice Suggests gam License GPL (>= 2) Repository CRAN Date/Publication 2010-01-21 06:23:21

R topics documented:
aic . . . . aicplot . . ais . . . . ang . . . . bad . . . . border . . chemdiab claw54 . . cldem . . cltest . . . cltrain . . co2 . . . . cp . . . . cpar . . . cpplot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 4 5 6 7 7 8 8 9 9 10 10 11 11 12

2 crit . . . . . . dat . . . . . . density.lf . . . diab . . . . . ethanol . . . . expit . . . . . tted.loct . . formula.loct gam.lf . . . . gam.slist . . . gcv . . . . . . gcvplot . . . geyser . . . . geyser.round . hatmatrix . . heart . . . . . insect . . . . iris . . . . . . kangaroo . . kappa0 . . . . kdeb . . . . . km.mrl . . . . lcv . . . . . . lcvplot . . . . left . . . . . . lf . . . . . . . lfeval . . . . lfgrid . . . . lfknots . . . . lim . . . . . lfmarg . . . . lines.loct . . livmet . . . . loct . . . . . loct.censor . loct.matrix . loct.quasi . . loct.raw . . loct.robust . lp . . . . . . lscv . . . . . lscv.exact . . lscvplot . . . mcyc . . . . . mine . . . . . mmsamp . . . morths . . . . none . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

R topics documented: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 14 14 15 16 16 17 18 18 19 19 20 20 21 22 22 23 23 24 25 26 27 28 28 29 30 30 31 32 32 33 33 34 35 36 37 38 39 41 42 43 44 44 45 46 46 47 47

R topics documented: penny . . . . . . . . . plot.eval . . . . . . . . plot.gcvplot . . . . . . plot.lfeval . . . . . . . plot.loct . . . . . . . plot.loct.1d . . . . . . plot.loct.2d . . . . . . plot.loct.3d . . . . . . plot.preplot.loct . . . plot.scb . . . . . . . . plotbyfactor . . . . . . points.loct . . . . . . predict.loct . . . . . . preplot.loct . . . . . . preplot.loct.raw . . . print.gcvplot . . . . . . print.lfeval . . . . . . . print.loct . . . . . . . print.preplot.loct . . . print.scb . . . . . . . . print.summary.loct . . rbox . . . . . . . . . . regband . . . . . . . . residuals.loct . . . . . right . . . . . . . . . . rv . . . . . . . . . . . rva . . . . . . . . . . . scb . . . . . . . . . . . sjpi . . . . . . . . . . . smooth.lf . . . . . . . spence.15 . . . . . . . spence.21 . . . . . . . spencer . . . . . . . . stamp . . . . . . . . . store . . . . . . . . . . summary.gcvplot . . . summary.loct . . . . summary.preplot.loct trimod . . . . . . . . . xbar . . . . . . . . . . Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3 48 48 49 50 50 52 52 53 54 54 55 56 56 57 58 59 60 60 61 61 62 62 63 64 64 65 66 66 67 68 69 70 71 71 72 72 73 74 74 75 76

aicplot

aic

Compute Akaikes Information Criterion.

Description The calling sequence for aic matches those for the locfit or locfit.raw functions. The t is not returned; instead, the returned object contains Akaikes information criterion for the t. The denition of AIC used here is -2*log-likelihood + pen*(tted d.f.). For quasi-likelihood, and local regression, this assumes the scale parameter is one. Other scale parameters can effectively be used by changing the penalty. The AIC score is exact (up to numerical roundoff) if the ev="data" argument is provided. Otherwise, the residual sum-of-squares and degrees of freedom are computed using locts standard interpolation based approximations. Usage aic(x, ..., pen=2) Arguments x ... pen See Also locfit, locfit.raw, aicplot model formula other arguments to loct penalty for the degrees of freedom term

aicplot

Compute an AIC plot.

Description The aicplot function loops through calls to the aic function (and hence to locfit), using a different smoothing parameter for each call. The returned structure contains the AIC statistic for each t, and can be used to produce an AIC plot. Usage aicplot(..., alpha)

ais Arguments ... alpha arguments to the aic, locfit functions.

Matrix of smoothing parameters. The aicplot function loops through calls to aic, using each row of alpha as the smoothing parameter in turn. If alpha is provided as a vector, it will be converted to a one-column matrix, thus interpreting each component as a nearest neighbor smoothing parameter.

Value An object with class "gcvplot", containing the smoothing parameters and AIC scores. The actual plot is produced using plot.gcvplot. See Also locfit, locfit.raw, gcv, aic, plot.gcvplot Examples
data(morths) plot(aicplot(deaths~age,weights=n,data=morths,family="binomial", alpha=seq( .2,1. ,by= . 5)))

ais

Australian Institute of Sport Dataset

Description The rst two columns are the gender of the athlete and their sport. The remaining 11 columns are various measurements made on the athletes. Usage data(ais) Format A dataframe. Source Cook and Weisberg (1994). References Cook and Weisberg (1994). An Introduction to Regression Graphics. Wiley, New York.

ang

ang

Angular Term for a Loct model.

Description The ang() function is used in a loct model formula to specify that a variable should be treated as an angular or periodic term. The scale argument is used to set the period. ang(x) is equivalent to lp(x,style="ang"). Usage ang(x,...) Arguments x ... References Loader, C. (1999). Local Regression and Likelihood. Springer, NY (Section 6.2). See Also locfit. Examples
# generate an x variable, and a response with period x <- seq( ,1,length=2 ) y <- sin(1 *pi*x)+rnorm(2 )/5 .2

numeric variable to be treated periodically. Other arguments to lp.

# compute the periodic local fit. Note the scale argument is period/(2pi) fit <- locfit(y~ang(x,scale= .2/(2*pi))) # plot the fit over a single period plot(fit) # plot the fit over the full range of the data plot(fit,xlim=c( ,1))

bad

bad

Example dataset for bandwidth selection

Description Example dataset from Loader (1999). Usage data(bad) Format Data Frame with x and y variables. References Loader, C. (1999). Bandwidth Selection: Classical or Plug-in? Annals of Statistics 27.

border

Cricket Batting Dataset

Description Scores in 265 innings for Australian batsman Allan Border. Usage data(border) Format A dataframe with day (decimalized); not out indicator and score. The not out indicator should be used as a censoring variable. Source Compiled from the Cricinfo archives. References CricInfo: The Home of Cricket on the Internet. http://www.cricinfo.com/

claw54

chemdiab

Chemical Diabetes Dataset

Description Numeric variables are rw, fpg, ga, ina and sspg. Classier cc is the Diabetic type. Usage data(chemdiab) Format Data frame with ve numeric measurements and categroical response. Source Reaven and Miller (1979). References Reaven, G. M. and Miller, R. G. (1979). An attempt to dene the nature of chemical diabetes using a multidimensional analysis. Diabetologia 16, 17-24.

claw54

Claw Dataset

Description A random sample of size 54 from the claw density of Marron and Wand (1992), as used in Figure 10.5 of Loader (1999). Usage data(claw54) Format Numeric vector with length 54. Source Randomly generated.

cldem References Loader, C. (1999). Local Regression and Likelihood. Springer, New York.

Marron, J. S. and Wand, M. P. (1992). Exact mean integrated squared error. Annals of Statistics 20, 712-736.

cldem

Example data set for classication

Description Observations from Figure 8.7 of Loader (1999). Usage data(cldem) Format Data Frame with x and y variables. References Loader, C. (1999). Local Regression and Likelihood. Springer, New York.

cltest

Test dataset for classication

Description 200 observations from a 2 population model. Under population 0, x1,i has a standard normal distribution, and x2,i = (2 x2 + zi )/3, where zi is also standard normal. Under population 1, 1,i x2,i = (2 x2 + zi )/3. The optimal classication regions form a checkerboard pattern, with 1,i horizontal boundary at x2 = 0, vertical boundaries at x1 = 2. This is the same model as the cltrain dataset. Usage data(cltest) Format Data Frame. Three variables x1, x2 and y. The latter indicates class membership.

10

co2

cltrain

Training dataset for classication

Description 200 observations from a 2 population model. Under population 0, x1,i has a standard normal distribution, and x2,i = (2 x2 + zi )/3, where zi is also standard normal. Under population 1, 1,i x2,i = (2 x2 + zi )/3. The optimal classication regions form a checkerboard pattern, with 1,i horizontal boundary at x2 = 0, vertical boundaries at x1 = 2. This is the same model as the cltest dataset. Usage data(cltrain) Format Data Frame. Three variables x1, x2 and y. The latter indicates class membership.

co2

Carbon Dioxide Dataset

Description Monthly time series of carbon dioxide measurements at Mauna Loa, Hawaii from 1959 to 1990. Usage data(co2) Format Data frame with year, month and co2 variables. Source Boden, Sepanski and Stoss (1992). References Boden, Sepanski and Stoss (1992). Trends 91: A compedium of data on global change - Highlights. Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory.

cp

11

cp

Compute Mallows Cp for local regression models.

Description The calling sequence for cp matches those for the locfit or locfit.raw functions. The t is not returned; instead, the returned object contains Cp criterion for the t. Cp is usually computed using a variance estimate from the largest model under consideration, rather than 2 = 1. This will be done automatically when the cpplot function is used. The Cp score is exact (up to numerical roundoff) if the ev="data" argument is provided. Otherwise, the residual sum-of-squares and degrees of freedom are computed using locts standard interpolation based approximations. Usage cp(x, ..., sig2=1) Arguments x ... sig2 See Also locfit, locfit.raw, cpplot model formula or numeric vector of the independent variable. other arguments to locfit and/or locfit.raw. residual variance estimate.

cpar

Conditionally parametric term for a Loct model.

Description A term entered in a locfit model formula using cpar will result in a t that is conditionally parametric. Equivalent to lp(x,style="cpar"). This function is presently almost deprecated. Specifying a conditionally parametric t as y~x1+cpar(x2) wil no longer work; instead, the model is specied as y~lp(x1,x2,style=c("n","cpar")). Usage cpar(x,...) Arguments x ... numeric variable. Other arguments to link{lp}().

12 See Also locfit Examples

data(ethanol, package="locfit") # fit a conditionally parametric model fit <- locfit(NOx ~ lp(E, C, style=c("n","cpar")), data=ethanol) plot(fit) # one way to force a parametric fit with locfit fit <- locfit(NOx ~ cpar(E), data=ethanol)

cpplot

cpplot

Compute a Cp plot.

Description The cpplot function loops through calls to the cp function (and hence to link{locfit}), using a different smoothing parameter for each call. The returned structure contains the Cp statistic for each t, and can be used to produce an AIC plot. Usage cpplot(..., alpha, sig2) Arguments ... alpha arguments to the cp, locfit functions. Matrix of smoothing parameters. The cpplot function loops through calls to cp, using each row of alpha as the smoothing parameter in turn. If alpha is provided as a vector, it will be converted to a one-column matrix, thus interpreting each component as a nearest neighbor smoothing parameter. Residual variance. If not specied, the residual variance is computed using the tted model with the fewest residual degrees of freedom.

sig2

Value An object with class "gcvplot", containing the smoothing parameters and CP scores. The actual plot is produced using plot.gcvplot. See Also locfit, locfit.raw, gcv, aic, plot.gcvplot Examples
data(ethanol) plot(cpplot(NOx~E,data=ethanol,alpha=seq( .2,1. ,by= . 5)))

crit

13

crit

Compute critical values for condence intervals.

Description Every "locfit" object contains a critical value object to be used in computing and ploting condence intervals. By default, a 95% pointwise condence level is used. To change the condence level, the critical value object must be substituted using crit and crit<-. Usage crit(fit, const=c( , 1), d=1, cov= .95, rdf= ) crit(fit) <- value Arguments fit "locfit" object. This is optional; if a t is provided, defaults for the other arguments are taken from the critical value currently stored on this t, rather than the usual values above. crit(fit) with no other arguments will just return the current critical value. Tube formula constants for simultaneous bands (the default, c( ,1), produces pointwise coverage). Usually this is generated by the kappa function and should not be provided by the user. Dimension of the t. Again, users shouldnt usually provide it. Coverage Probability for critical values. Residual degrees of freedom. If non-zero, the critical values are based on the Students t distribution. When rdf= , the normal distribution is used. Critical value object generated by crit or kappa .

const

d cov rdf value Value

Critical value object. See Also locfit, plot.locfit, kappa , crit<-. Examples
# compute and plot 99% confidence intervals, with local variance estimate. data(ethanol) fit <- locfit(NOx~E,data=ethanol) crit(fit) <- crit(fit,cov= .99) plot(fit,band="local") # compute and plot 99% simultaneous bands crit(fit) <- kappa (NOx~E,data=ethanol,cov= .99) plot(fit,band="local")

14

density.lf

dat

Loct - data evaluation structure.

Description dat is used to specify evaluation on the given data points for locfit.raw(). Usage dat(cv=FALSE) Arguments cv Whether cross-validation should be done.

density.lf

Density estimation using Loct

Description This function provides an interface to Loct, in the syntax of (a now old version of) the S-Plus density function. This can reproduce density results, but allows additional locfit.raw arguments, such as the degree of t, to be given. It also works in double precision, whereas density only works in single precision. Usage density.lf(x, n = 5 , window = "gaussian", width, from, to, cut = if(iwindow == 4.) .75 else .5, ev = lfgrid(mg = n, ll = from, ur = to), deg = , family = "density", link = "ident", ...) Arguments x n window width from to cut numeric vector of observations whose density is to be estimated. number of evaluation points. Equivalent to the locfit.raw mg argument. Window type to use for estimation. Equivalent to the locfit.raw kern argument. This includes all the density windows except cosine. Window width. Following density, this is the full width; not the half-width usually used by Loct and many other smoothers. Lower limit for estimation domain. Upper limit for estimation domain. Controls default expansion of the domain.

diab ev deg family link ... Value A list with components x (evaluation points) and y (estimated density). See Also density, locfit, locfit.raw Examples
data(geyser) density.lf(geyser, window="tria") # the same result with density, except less precision. density(geyser, window="tria")

15 Loct evaluation structure default lfgrid(). Fitting degree default 0 for kernel estimation. Fitting family default is "density". Link function default is the "identity". Additional arguments to locfit.raw, with standard defaults.

diab

Exhaust emissions

Description NOx exhaust emissions from a single cylinder engine. Two predictor variables are E (the engines equivalence ratio) and C (Compression ratio). Usage data(ethanol) Format Data frame with NOx, E and C variables. Source Brinkman (1981). Also studied extensively by Cleveland (1993). References Brinkman, N. D. (1981). Ethanol fuel - a single-cylinder engine study of efciency and exhaust emissions. SAE transactions 90, 1414-1424. Cleveland, W. S. (1993). Visualizing data. Hobart Press, Summit, NJ.

16

expit

ethanol

Exhaust emissions

Description NOx exhaust emissions from a single cylinder engine. Two predictor variables are E (the engines equivalence ratio) and C (Compression ratio). Usage data(ethanol) Format Data frame with NOx, E and C variables. Source Brinkman (1981). Also studied extensively by Cleveland (1993). References Brinkman, N. D. (1981). Ethanol fuel - a single-cylinder engine study of efciency and exhaust emissions. SAE transactions 90, 1414-1424. Cleveland, W. S. (1993). Visualizing data. Hobart Press, Summit, NJ.

expit

Inverse logistic link function

Description Computes ex /(1 + ex ). This is the inverse of the logistic link function, log(p/(1 p)). Usage expit(x) Arguments x numeric vector

tted.loct

17

fitted.locfit

Fitted values for a "loct" object.

Description Evaluates the tted values (i.e. evaluates the surface at the original data points) for a Loct object. This function works by reconstructing the model matrix from the original formula, and predicting at those points. The function may be fooled; for example, if the original data frame has changed since the t, or if the model formula includes calls to random number generators. Usage ## S3 method for class locfit fitted(object, data=NULL, what="coef", cv=FALSE, studentize=FALSE, type="fit", tr, ...) Arguments object data "locfit" object. The data frame for the original t. Usually, this shouldnt be needed, especially when the function is called directly. It may be needed when called inside another function. What to compute tted values of. The default, what="coef", works with the tted curve itself. Other choices include "nlx" for the length of the weight diagram; "infl" for the inuence function; "band" for the bandwidth; "degr" for the local polynomial degree; "lik" for the maximized local likelihood; "rdf" for the local residual degrees of freedom and "vari" for the variance function. The interpolation algorithm for some of these quantities is questionable. If TRUE, leave-one-out cross validated tted values are approximated. Wont make much sense, unless what="coef". If TRUE, residuals are studentized. Type of t or residuals to compute. The default is "fit" for fitted.locfit, and "dev" for residuals.locfit. Other choices include "pear" for Pearson residuals; "raw" for raw residuals, "ldot" for likelihood derivative; "d2" for the deviance residual squared; lddot for the likelihood second derivative. Generally, type should only be used when what="coef". Back transformation for likelihood models. arguments passed to and from methods.

what

cv studentize type

tr ... Value

A numeric vector of the tted values. See Also locfit, predict.locfit, residuals.locfit

18

gam.lf

formula.locfit

Formula from a Loct object.

Description Extract the model formula from a loct object. Usage ## S3 method for class locfit formula(x, ...) Arguments x ... Value Returns the formula from the loct object. See Also locfit locfit object. Arguments passed to and from other methods.

gam.lf

Loct call for Generalized Additive Models

Description This is a loct calling function used by lf() terms in additive models. It is not normally called directly by users. Usage gam.lf(x, y, w, xeval, ...) Arguments x y w xeval ... numeric predictor numeric response prior weights evaluation points other arguments to locfit.raw()

gam.slist See Also locfit, locfit.raw, lf, gam

19

gam.slist

Vector of GAM special terms

Description This vector adds "lf" to the default vector of special terms recognized by a gam() model formula. To ensure this is recognized, attach the Loct library with library(locfit,first=T). Format Character vector. See Also lf, gam

gcv

Compute generalized cross-validation statistic.

Description The calling sequence for gcv matches those for the locfit or locfit.raw functions. The t is not returned; instead, the returned object contains Wahbas generalized cross-validation score for the t. The GCV score is exact (up to numerical roundoff) if the ev="data" argument is provided. Otherwise, the residual sum-of-squares and degrees of freedom are computed using locts standard interpolation based approximations. For likelihood models, GCV is computed uses the deviance in place of the residual sum of squares. This produces useful results but I do not know of any theory validating this extension. Usage gcv(x, ...) Arguments x, ... See Also locfit, locfit.raw, gcvplot Arguments passed on to locfit or locfit.raw.

20

geyser

gcvplot

Compute a generalized cross-validation plot.

Description The gcvplot function loops through calls to the gcv function (and hence to link{locfit}), using a different smoothing parameter for each call. The returned structure contains the GCV statistic for each t, and can be used to produce an GCV plot. Usage gcvplot(..., alpha, df=2) Arguments ... alpha arguments to the gcv, locfit functions. Matrix of smoothing parameters. The gcvplot function loops through calls to gcv, using each row of alpha as the smoothing parameter in turn. If alpha is provided as a vector, it will be converted to a one-column matrix, thus interpreting each component as a nearest neighbor smoothing parameter. Degrees of freedom to use as the x-axis. 2=trace(L), 3=trace(LL).

df Value

An object with class "gcvplot", containing the smoothing parameters and GCV scores. The actual plot is produced using plot.gcvplot. See Also locfit, locfit.raw, gcv, plot.gcvplot, summary.gcvplot Examples
data(ethanol) plot(gcvplot(NOx~E,data=ethanol,alpha=seq( .2,1. ,by= . 5)))

geyser

Old Faithful Geyser Dataset

Description The durations of 107 eruptions of the Old Faithful Geyser. Usage data(geyser)

geyser.round Format A numeric vector of length 107. Source

21

Scott (1992). Note that several different Old Faithful Geyser datasets (including the faithful dataset in Rs base library) have been used in various places in the statistics literature. The version provided here has been used in density estimation and bandwidth selection work. References Scott, D. W. (1992). Multivariate Density Estimation: Theory, Practice and Visualization. Wiley.

geyser.round

Discrete Old Faithful Geyser Dataset

Description This is a variant of the geyser dataset, where each observation is rounded to the nearest 0.05 minutes, and the counts tallied. Usage data(geyser.round) Format Data Frame with variables duration and count. Source Scott (1992). Note that several different Old Faithful Geyser datasets (including the faithful dataset in Rs base library) have been used in various places in the statistics literature. The version provided here has been used in density estimation and bandwidth selection work. References Scott, D. W. (1992). Multivariate Density Estimation: Theory, Practice and Visualization. Wiley.

22

heart

hatmatrix

Weight diagrams and the hat matrix for a local regression model.

Description hatmatrix() computes the weight diagrams (also known as equivalent or effective kernels) for a local regression smooth. Essentially, hatmatrix() is a front-end to locfit(), setting a ag to compute and return weight diagrams, rather than the t. Usage hatmatrix(formula, dc=TRUE, ...) Arguments formula dc ... Value A matrix with n rows and p columns; each column being the weight diagram for the corresponding locfit t point. If ev="data", this is the transpose of the hat matrix. See Also locfit, plot.locfit.1d, plot.locfit.2d, plot.locfit.3d, lines.locfit, predict.locfit model formula. derivative adjustment (see locfit.raw) Other arguments to locfit and locfit.raw.

heart

Survival Times of Heart Transplant Recipients

Description The survival times of 184 participants in the Stanford heart transplant program. Usage data(heart) Format Data frame with surv, cens and age variables.

insect Source

23

Miller and Halperin (1982). The original dataset includes information on additional patients who never received a transplant. Other authors reported earlier versions of the data. References Miller, R. G. and Halperin, J. (1982). Regression with censored data. Biometrika 69, 521-531.

insect

Insect Dataset

Description An experiment measuring death rates for insects, with 30 insects at each of ve treatment levels. Usage data(insect) Format Data frame with lconc (dosage), deaths (number of deaths) and nins (number of insects) variables. Source Bliss (1935). References Bliss (1935). The calculation of the dosage-mortality curve. Annals of Applied Biology 22, 134-167.

iris

Fishers Iris Data (subset)

Description Four measurements on each of fty owers of two species of iris (Versicolor and Virginica) A classication dataset. Fishers original dataset contained a third species (Setosa) which is trivially seperable. Usage data(iris)

24 Format Data frame with species, petal.wid, petal.len, sepal.wid, sepal.len.

kangaroo

Source Fisher (1936). Reproduced in Andrews and Herzberg (1985) Chapter 1.

References Andrews, D. F. and Herzberg, A. M. (1985). Data. Springer-Verlag. Fisher, R. A. (1936). The Use of Multiple Measurements in Taxonomic Problems. Annals of Eugenics 7, Part II. 179-188.

kangaroo

Kangaroo skull measurements dataset

Description Variables are sex (m/f), spec (giganteus, melanops, fuliginosus) and 18 numeric measurements.

Usage data(kangaroo)

Format Data frame with measurements on the skulls of 101 kangaroos. (number of insects) variables.

Source Andrews and Herzberg (1985) Chapter 53.

References Andrews, D. F. and Herzberg, A. M. (1985). Data. Springer-Verlag, New York.

kappa0

25

kappa

Critical Values for Simultaneous Condence Bands.

Description The geometric constants for simultaneous condence bands are computed, as described in Sun and Loader (1994) (bias adjustment is not implemented here). These are then passed to the crit function, which computes the critical value for the condence bands. The method requires both the weight diagrams l(x), the derivative l(x) and (in 2 or more dimensions) the second derivatives l(x). These are implemented exactly for a constant bandwidth. For nearest neighbor bandwidths, the computations are approximate and a warning is produced. The theoretical justication for the bands uses normality of the random errors e1 , . . . , en in the regression model, and in particular the spherical symmetry of the error vector. For non-normal distributions, and likelihood models, one relies on central limit and related theorems. Computation uses the product Simpsons rule to evaluate the multidimensional integrals (The domain of integration, and hence the region of simultaneous coverage, is determined by the flim argument). Expect the integration to be slow in more than one dimension. The mint argument controls the precision. Usage kappa (formula, cov= .95, ev=lfgrid(2 ), ...) Arguments formula cov ev ... Value A list with components for the critical value, geometric constants, e.t.c. Can be passed directly to plot.locfit as the crit argument. References Sun, J. and Loader, C. (1994). Simultaneous condence bands for linear regression and smoothing. Annals of Statistics 22, 1328-1345. See Also locfit, plot.locfit, crit, crit<-. Local regression model formula. A "locfit" object can also be provided; in this case the formula and other arguments are extracted from this object. Coverage Probability for critical values. Loct evaluation structure. Should usually be a grid this species the integration rule. Other arguments to locfit. Important arguments include flim and alpha.

26 Examples
# compute and plot simultaneous confidence bands data(ethanol) fit <- locfit(NOx~E,data=ethanol) crit(fit) <- kappa (NOx~E,data=ethanol) plot(fit,crit=crit,band="local")

kdeb

kdeb

Bandwidth selectors for kernel density estimation.

Description Function to compute kernel density estimate bandwidths, as used in the simulation results in Chapter 10 of Loader (1999). This function is included for comparative purposes only. Plug-in selectors are based on awed logic, make unreasonable and restrictive assumptions and do not use the full power of the estimates available in Loct. Any relation between the results produced by this function and desirable estimates are entirely coincidental. Usage kdeb(x, h = . 1 * sd, h1 = sd, meth = c("AIC", "LCV", "LSCV", "BCV", "SJPI", "GKK"), kern = "gauss", gf = 2.5)

Arguments x h h1 meth kern gf One dimensional data vector. Lower limit for bandwidth selection. Can be fairly small, but h0=0 would cause problems. Upper limit. Required selection method(s). Kernel. Most methods require kern="gauss", the default for this function only. Standard deviation for the gaussian kernel. Default 2.5, as Locts standard. Most papers use 1.

Value Vector of selected bandwidths. References Loader, C. (1999). Local Regression and Likelihood. Springer, New York.

km.mrl

27

km.mrl

Mean Residual Life using Kaplan-Meier estimate

Description This function computes the mean residual life for censored data using the Kaplan-Meier estimate of the survival function. If S(t) is the K-M estimate, the MRL for a censored observation is computed as ( t S(u)du)/S(t). We take S(t) = 0 when t is greater than the largest observation, regardless of whether that observation was censored. When there are ties between censored and uncensored observations, for deniteness our ordering places the censored observations before uncensored. This function is used by locfit.censor to compute censored regression estimates. Usage km.mrl(times, cens) Arguments times cens Obsereved survival times. Logical variable indicating censoring. The coding is 1 or TRUE for censored; or FALSE for uncensored.

Value A vector of the estimated mean residual life. For uncensored observations, the corresponding estimate is 0. References Buckley, J. and James, I. (1979). Linear Regression with censored data. Biometrika 66, 429-436. Loader, C. (1999). Local Regression and Likelihood. Springer, NY (Section 7.2). See Also locfit.censor Examples
# censored regression using the Kaplan-Meier estimate. data(heart, package="locfit") fit <- locfit.censor(log1 (surv+ .5)~age, cens=cens, data=heart, km=TRUE) plotbyfactor(heart$age, .5+heart$surv, heart$cens, ylim=c( .5,16 ), log="y") lines(fit, tr=function(x)1 ^x)

28

lcvplot

lcv

Compute Likelihood Cross Validation Statistic.

Description The calling sequence for lcv matches those for the locfit or locfit.raw functions. The t is not returned; instead, the returned object contains likelihood cross validation score for the t. The LCV score is exact (up to numerical roundoff) if the ev="cross" argument is provided. Otherwise, the inuence and cross validated residuals are computed using locts standard interpolation based approximations. Usage lcv(x, ...) Arguments x ... See Also locfit, locfit.raw, lcvplot model formula other arguments to loct

lcvplot

Compute the likelihood cross-validation plot.

Description The lcvplot function loops through calls to the lcv function (and hence to link{locfit}), using a different smoothing parameter for each call. The returned structure contains the likelihood cross validation statistic for each t, and can be used to produce an LCV plot. Usage lcvplot(..., alpha) Arguments ... alpha arguments to the lcv, locfit functions. Matrix of smoothing parameters. The aicplot function loops through calls to lcv, using each row of alpha as the smoothing parameter in turn. If alpha is provided as a vector, it will be converted to a one-column matrix, thus interpreting each component as a nearest neighbor smoothing parameter.

left Value

29

An object with class "gcvplot", containing the smoothing parameters and LCV scores. The actual plot is produced using plot.gcvplot. See Also locfit, locfit.raw, gcv, lcv, plot.gcvplot Examples
data(ethanol) plot(lcvplot(NOx~E,data=ethanol,alpha=seq( .2,1. ,by= . 5)))

left

One-sided left smooth for a Loct model.

Description The left() function is used in a loct model formula to specify a one-sided smooth: when tting at a point x, only data points with xi x should be used. This can be useful in estimating points of discontinuity, and in cross-validation for forecasting a time series. left(x) is equivalent to lp(x,style="left"). When using this function, it will usually be necessary to specify an evaluation structure, since the t is not smooth and locts interpolation methods are unreliable. Also, it is usually best to use deg= or deg=1, otherwise the ts may be too variable. If nearest neighbor bandwidth specication is used, it does not recognize left(). Usage left(x,...) Arguments x ... See Also locfit, lp, right Examples
# compute left and right smooths data(penny) xev <- (1945:1988)+ .5 fitl <- locfit(thickness~left(year,h=1 ,deg=1), ev=xev, data=penny) fitr <- locfit(thickness~right(year,h=1 ,deg=1),ev=xev, data=penny) # plot the squared difference, to show the change points. plot( xev, (predict(fitr,where="ev") - predict(fitl,where="ev"))^2 )

numeric variable. Other arguments to lp().

30

lfeval

lf

Loct term in Additive Model formula

Description This function is used to specify a smooth term in a gam() model formula. This function is designed to be used with the S-Plus gam() function. For R users, there are at least two different gam() functions available. Most current distributions of R will include the mgcv library by Simon Wood; lf() is not compatable with this function. On CRAN, there is a gam package by Trevor Hastie, similar to the S-Plus version. lf() should be compatable with this, although its untested. Usage lf(..., alpha= .7, deg=2, scale=1, kern="tcub", ev=rbox(), maxk=1 ) Arguments ... numeric predictor variable(s) alpha, deg, scale, kern, ev, maxk these are as in locfit.raw. See Also locfit, locfit.raw, gam.lf, gam Examples
# fit an additive semiparametric model to the ethanol data. stopifnot(require(gam)) # The gam package must be attached _before_ locfit, otherwise # the following will not work. fit <- gam(NOx ~ lf(E) + C, data=ethanol) op <- par(mfrow=c(2, 1)) plot(fit) par(op)

lfeval

Extract Loct Evaluation Structure.

Description Extracts the evaluation structure from a "locfit" object. This object has the class "lfeval", and has its own set of methods for plotting e.t.c.

lfgrid Usage lfeval(object) Arguments object Value "lfeval" object. See Also locfit, plot.lfeval, print.lfeval "locfit" object

31

lfgrid

Loct - grid evaluation structure.

Description lfgrid() is used to specify evaluation on a grid of points for locfit.raw(). The structure computes a bounding box for the data, and divides that into a grid with specied margins. Usage lfgrid(mg=1 , ll, ur) Arguments mg ll ur Number of grid points along each margin. Can be a single number (which is applied in each dimension), or a vector specifying a value for each dimension. Lower left limits for the grid. Length should be the number of dimensions of the data provided to locfit.raw(). Upper right limits for the grid. By default, ll and ur are generated as the bounding box for the data.

Examples
data(ethanol, package="locfit") plot.eval(locfit(NOx ~ lp(E, C, scale=TRUE), data=ethanol, ev=lfgrid()))

32

lim

lfknots

Extraction of t-point information from a Loct object.

Description Extracts information, such as tted values, inuence functions from a "locfit" object. Usage lfknots(x, tr, what = c("x", "coef", "h", "nlx"), delete.pv = TRUE) Arguments x tr what Fitted object from locfit(). Back transformation. Default is the invers link function from the Loct object. What to return; default is c("x","coef","h","nlx"). Allowed elds are x (t points); coef (tted values); f1 (local slope); nlx (length of the weight diagram); nlx1 (estimated derivative of nlx); se (standard errors); infl (inuence function); infla (slope of inuence function); lik (maximixed local log-likelihood and local degrees of freedom); h (bandwidth) and deg (degree of t). If T, pseudo-vertices are deleted.

delete.pv Value

A matrix with one row for each t point. Columns correspond to the specied what vector; some elds contribute multiple columns.

lflim

Construct Limit Vectors for Loct ts.

Description This function is used internally to interpret xlim and flim arguments. It should not be called directly. Usage lflim(limits, nm, ret) Arguments limits nm ret Limit argument. Variable names. Initial return vector.

lfmarg Value Vector with length 2*dim. See Also locfit

33

lfmarg

Generate grid margins.

Description This function is usually called by plot.locfit. Usage lfmarg(xlim, m = 4 ) Arguments xlim Vector of limits for the grid. Should be of length 2*d; the rst d components represent the lower left corner, and the next d components the upper right corner. Can also be a "locfit" object. Number of points for each grid margin. Can be a vector of length d.

m Value

A list, whose components are the d grid margins. See Also locfit, plot.locfit

lines.locfit

Add loct line to existing plot

Description Adds a Loct line to an existing plot. llines is for use within a panel function for Lattice. Usage ## S3 method for class locfit lines(x, m=1 , tr=x$trans, ...) ## S3 method for class locfit llines(x, m=1 , tr=x$trans, ...)

34 Arguments x m tr ... locfit object. Should be a model with one predictor. Number of points to evaluate the line at.

livmet

Transformation function to use for plotting. Default is the inverse link function, or the identity function if derivatives are required. Other arguments to the default lines function.

See Also locfit, plot.locfit, lines

livmet

liver Metastases dataset

Description Survival times for 622 patients diagnosed with Liver Metastases. Beware, the censoring variable is coded as 1 = uncensored, so use cens=1-z in locfit() calls.

Usage data(livmet)

Format Data frame with survival times (t), censoring indicator (z) and a number of covariates.

Source Haupt and Mansmann (1995)

References Haupt, G. and Mansmann, U. (1995) CART for Survival Data. Statlib Archive, http://lib.stat. cmu.edu/S/survcart.

loct

35

locfit

Local Regression, Likelihood and Density Estimation.

Description locfit is the model formula-based interface to the Loct library for tting local regression and likelihood models. locfit is implemented as a front-end to locfit.raw. See that function for options to control smoothing parameters, tting family and other aspects of the t. Usage locfit(formula, data=sys.frame(sys.parent()), weights=1, cens= , base= , subset, geth=FALSE, ..., lfproc=locfit.raw) Arguments formula Model Formula; e.g. y~lp(x) for a regression model; ~lp(x) for a density estimation model. Use of lp() on the RHS is recommended, especially when non-default smoothing parameters are used. Data Frame. Prior weights (or sample sizes) for individual observations. This is typically used where observations have unequal variance. Censoring indicator. 1 (or TRUE) denotes a censored observation. denotes uncensored. (or FALSE)

data weights cens base

Baseline for local tting. For local regression models, specifying a base is equivalent to using y-base as the reponse. But base also works for local likelihood. Subset observations in the data frame. Dont use. Other arguments to locfit.raw() (or the lfproc). A processing function to compute the local t. Default is locfit.raw(). Other choices include locfit.robust(), locfit.censor() and locfit.quasi().

subset geth ... lfproc

Value An object with class "locfit". A standard set of methods for printing, ploting, etc. these objects is provided. References Loader, C. (1999). Local Regression and Likelihood. Springer, New York.

36 See Also locfit.raw Examples

# fit and plot a univariate local regression data(ethanol, package="locfit") fit <- locfit(NOx ~ E, data=ethanol) plot(fit, get.data=TRUE) # a bivariate local regression with smaller smoothing parameter fit <- locfit(NOx~lp(E,C,nn= .5,scale= ), data=ethanol) plot(fit) # density estimation data(geyser, data="locfit") fit <- locfit( ~ lp(geyser, nn= .1, h= .8)) plot(fit,get.data=TRUE)

loct.censor

locfit.censor

Censored Local Regression

Description locfit.censor produces local regression estimates for censored data. The basic idea is to use an EM style algorithm, where one alternates between estimating the regression and the true values of censored observations. locfit.censor is designed as a front end to locfit.raw with data vectors, or as an intemediary between locfit and locfit.raw with a model formula. If you can stand the syntax, the second calling sequence above will be slightly more efcient than the third. Usage locfit.censor(x, y, cens, ..., iter=3, km=FALSE) Arguments x y cens ... iter km Either a locfit model formula or a numeric vector of the predictor variable. If x is numeric, y gives the response variable. Logical variable indicating censoring. The coding is 1 or TRUE for censored; or FALSE for uncensored. Other arguments to locfit.raw Number of EM iterations to perform If km=TRUE, the estimation of censored observations uses the Kaplan-Meier estimate, leading to a local version of the Buckley-James estimate. If km=F, the estimation is based on a normal model (Schmee and Hahn). Beware of claims that B-J is nonparametric; it makes stronger assumptions on the upper tail of survival distributions than most authors care to admit.

loct.matrix Value locfit object. References

37

Buckley, J. and James, I. (1979). Linear Regression with censored data. Biometrika 66, 429-436. Loader, C. (1999). Local Regression and Likelihood. Springer, NY (Section 7.2). Schmee, J. and Hahn, G. J. (1979). A simple method for linear regression analysis with censored data (with discussion). Technometrics 21, 417-434. See Also km.mrl, locfit, locfit.raw Examples
data(heart, package="locfit") fit <- locfit.censor(log1 (surv+ .5) ~ age, cens=cens, data=heart) ## Can also be written as: ## Not run: fit <- locfit(log1 (surv + .5) ~ age, cens=cens, data=heart, lfproc=locfit.censor) with(heart, plotbyfactor(age, .5 + surv, cens, ylim=c( .5, 16 ), log="y")) lines(fit, tr=function(x) 1 ^x)

locfit.matrix

Reconstruct a Loct model matrix.

Description Reconstructs the model matrix, and associated variables such as the response, prior weights and censoring indicators, from a locfit object. This is used by functions such as fitted.locfit; it is not normally called directly. The function will only work properly if the data frame has not been changed since the t was constructed. Usage locfit.matrix(fit, data) Arguments fit data Value A list with variables x (the model matrix); y (the response); w (prior weights); sc (scales); ce (censoring indicator) and base (baseline t). Loct object Data Frame.

38 See Also locfit, fitted.locfit, residuals.locfit

loct.quasi

locfit.quasi

Local Quasi-Likelihood with global reweighting.

Description locfit.quasi assumes a specied mean-variance relation, and performs iterartive reweighted local regression under this assumption. This is appropriate for local quasi-likelihood models, and is an alternative to specifying a family such as "qpoisson". locfit.quasi is designed as a front end to locfit.raw with data vectors, or as an intemediary between locfit and locfit.raw with a model formula. If you can stand the syntax, the second calling sequence above will be slightly more efcient than the third. Usage locfit.quasi(x, y, weights, ..., iter=3, var=abs) Arguments x y weights ... iter var Value "locfit" object. See Also locfit, locfit.raw Either a locfit model formula or a numeric vector of the predictor variable. If x is numeric, y gives the response variable. Case weights to use in the tting. Other arguments to locfit.raw Number of EM iterations to perform Function specifying the assumed relation between the mean and variance.

loct.raw

39

locfit.raw

Local Regression, Likelihood and Density Estimation.

Description locfit.raw is an interface to Loct using numeric vectors (for a model-formula based interface, use locfit). Although this function has a large number of arguments, most users are likely to need only a small subset. The rst set of arguments (x, y, weights, cens, and base) specify the regression variables and associated quantities. Another set (scale, alpha, deg, kern, kt, acri and basis) control the amount of smoothing: bandwidth, smoothing weights and the local model. Most of these arguments are deprecated theyll currently still work, but should be provided through the lp() model term instead. deriv and dc relate to derivative (or local slope) estimation. family and link specify the likelihood family. xlim and renorm may be used in density estimation. ev species the evaluation structure or set of evaluation points. maxk, itype, mint, maxit and debug control the Loct algorithms, and will be rarely used. geth and sty are used by other functions calling locfit.raw, and should not be used directly. Usage locfit.raw(x, y, weights=1, cens= , base= , scale=FALSE, alpha= .7, deg=2, kern="tricube", kt="sph", acri="none", basis=list(NULL), deriv=numeric( ), dc=FALSE, family, link="default", xlim, renorm=FALSE, ev=rbox(), maxk=1 , itype="default", mint=2 , maxit=2 , debug= , geth=FALSE, sty="none") Arguments x y weights cens base Vector (or matrix) of the independent variable(s). Can be constructed using the lp() function. Response variable for regression models. For density families, y can be omitted. Prior weights for observations (reciprocal of variance, or sample size). Censoring indicators for hazard rate or censored regression. The coding is 1 (or TRUE) for a censored observation, and (or FALSE) for uncensored observations. Baseline parameter estimate. If provided, the local regression model is tted as Yi = bi +m(xi )+ i , with Loct estimating the m(x) term. For regression models, this effectively subtracts bi from Yi . The advantage of the base formulation is that it extends to likelihood regression models.

40 scale alpha Deprecated - see lp().

loct.raw

Deprecated - see lp(). A single number (e.g. alpha= .7) is interpreted as a nearest neighbor fraction. With two componentes (e.g. alpha=c( .7,1.2)), the rst component is a nearest neighbor fraction, and the second component is a xed component. A third component is the penalty term in locally adaptive smoothing. Degree of local polynomial. Deprecated - see lp(). Weight function, default = "tcub". Other choices are "rect", "trwt", "tria", "epan", "bisq" and "gauss". Choices may be restricted when derivatives are required; e.g. for condence bands and some bandwidth selectors. Kernel type, "sph" (default); "prod". In multivariate problems, "prod" uses a simplied product model which speeds up computations. Deprecated - see lp(). User-specied basis functions. See lfbas for more details on this argument. Derivative estimation. If deriv=1, the returned t will be estimating the derivative (or more correctly, an estimate of the local slope). If deriv=c(1,1) the second order derivative is estimated. deriv=2 is for the partial derivative, with respect to the second variable, in multivariate settings. Derivative adjustment. Local likelihood family; "gaussian"; "binomial"; "poisson"; "gamma" and "geom". Density and rate estimation families are "dens", "rate" and "hazard" (hazard rate). If the family is preceded by a q (for example, family="qbinomial"), quasi-likelihood variance estimates are used. Otherwise, the residual variance (rv) is xed at 1. The default family is "qgauss" if a response y is provided; "density" if no response is provided. Link function for local likelihood tting. Depending on the family, choices may be "ident", "log", "logit", "inverse", "sqrt" and "arcsin". For density estimation, Loct allows the density to be supported on a bounded interval (or rectangle, in more than one dimension). The format should be c(ll,ul) where ll is a vector of the lower bounds and ur the upper bounds. Bounds such as [0, ) are not supported, but can be effectively implemented by specifying a very large upper bound. Local likelihood density estimates may not integrate exactly to 1. If renorm=T, the integral will be estimated numerically and the estimate rescaled. Presently this is implemented only in one dimension. The evaluation structure, rbox() for tree structures; lfgrid() for grids; dat() for data points; none() for none. A vector or matrix of evaluation points can also be provided, although in this case you may prefer to use the smooth.lf() interface to Loct. Note that arguments flim, mg and cut are now given as arguments to the evaluation structure function, rather than to locfit.raw() directly (change effective 12/2001). Controls space assignment for evaluation structures. For the adaptive evaluation structures, it is impossible to be sure in advance how many vertices will be generated. If you get warnings about Insufcient vertex space, Locts default assigment can be increased by increasing maxk. The default is maxk=1 .

deg kern

kt acri basis deriv

dc family

link xlim

renorm

ev

maxk

loct.robust itype

41 Integration type for density estimation. Available methods include "prod", "mult" and "mlin"; and "haz" for hazard rate estimation problems. The available integration methods depend on model specication (e.g. dimension, degree of t). By default, the best available method is used. Points for numerical integration rules. Default 20. Maximum iterations for local likelihood estimation. Default 20. If > 0; prints out some debugging information. Dont use! Deprecated - see lp().

mint maxit debug geth sty Value

An object with class "loct". A standard set of methods for printing, ploting, etc. these objects is provided. References Consult the Web page http://www.locfit.info/.

locfit.robust

Robust Local Regression

Description locfit.robust implements a robust local regression where outliers are iteratively identied and downweighted, similarly to the lowess method (Cleveland, 1979). The iterations and scale estimation are performed on a global basis. The scale estimate is 6 times the median absolute residual, while the robust downweighting uses the bisquare function. These are performed in the S code so easily changed. This can be interpreted as an extension of M estimation to local regression. An alternative extension (implemented in loct via family="qrgauss") performs the iteration and scale estimation on a local basis. Usage locfit.robust(x, y, weights, ..., iter=3) Arguments x y weights ... iter Either a locfit model formula or a numeric vector of the predictor variable. If x is numeric, y gives the response variable. weights to use in the tting. Other arguments to locfit.raw. Number of iterations to perform

42 Value "locfit" object. References

lp

Cleveland, W. S. (1979). Robust locally weighted regression and smoothing scatterplots. J. Amer. Statist. Assn. 74, 829-836. See Also locfit, locfit.raw

lp

Local Polynomial Model Term

Description lp is a local polynomial model term for Loct models. Usually, it will be the only term on the RHS of the model formula. Smoothing parameters should be provided as arguments to lp(), rather than to locfit(). Usage lp(..., nn, h, adpen, deg, acri, scale, style) Arguments ... nn h adpen deg acri style scale Predictor variables for the local regression model. Nearest neighbor component of the smoothing parameter. Default value is 0.7, unless either h or adpen are provided, in which case the default is 0. The constant component of the smoothing parameter. Default: 0. Penalty parameter for adaptive tting. Degree of polynomial to use. Criterion for adaptive bandwidth selection. Style for special terms (left, ang e.t.c.). Do not try to set this directly; call locfit instead. A scale to apply to each variable. This is especially important for multivariate tting, where variables may be measured in non-comparable units. It is also used to specify the frequency for ang terms. If scale=F (the default) no scaling is performed. If scale=T, marginal standard deviations are used. Alternatively, a numeric vector can provide scales for the individual variables.

See Also locfit, locfit.raw

lscv Examples
data(ethanol, package="locfit") # fit with 5 % nearest neighbor bandwidth. fit <- locfit(NOx~lp(E,nn= .5),data=ethanol) # bivariate fit. fit <- locfit(NOx~lp(E,C,scale=TRUE),data=ethanol) # density estimation data(geyser, package="locfit") fit <- locfit.raw(lp(geyser,nn= .1,h= .8))

43

lscv

Least Squares Cross Validation Statistic.

Description The calling sequence for lscv matches those for the locfit or locfit.raw functions. Note that this function is only designed for density estimation in one dimension. The returned object contains the least squares cross validation score for the t. The computation of f (x)2 dx is performed numerically. For kernel density estimation, this is unlikely to agree exactly with other LSCV routines, which may perform the integration analytically. Usage lscv(x, ..., exact=FALSE) Arguments x ... exact model formula (or numeric vector, if exact=T) other arguments to locfit or lscv.exact By default, the computation is approximate. If exact=TRUE, exact computation using lscv.exact is performed. This uses kernel density estimation with a constant bandwidth.

Value A vector consisting of the LSCV statistic and tted degrees of freedom. See Also locfit, locfit.raw, lscv.exact lscvplot Examples
# approximate calculation for a kernel density estimate data(geyser, package="locfit") lscv(~lp(geyser,h=1,deg= ), ev=lfgrid(1 ,ll=1,ur=6), kern="gauss") # same computation, exact lscv(lp(geyser,h=1),exact=TRUE)

44

lscvplot

lscv.exact

Exact LSCV Calculation

Description This function performs the exact computation of the least squares cross validation statistic for onedimensional kernel density estimation and a constant bandwidth. At the time of writing, it is implemented only for the Gaussian kernel (with the standard deviation of 0.4; Locts standard). Usage lscv.exact(x, h= ) Arguments x h Numeric data vector. The bandwidth. If x is constructed with lp(), the bandwidth should be given there instead.

Value A vector of the LSCV statistic and the tted degrees of freedom. See Also lscv, lscvplot Examples
data(geyser, package="locfit") lscv.exact(lp(geyser,h= .25)) # equivalent form using lscv lscv(lp(geyser, h= .25), exact=TRUE)

lscvplot

Compute the LSCV plot.

Description The lscvplot function loops through calls to the lscv function (and hence to link{locfit}), using a different smoothing parameter for each call. The returned structure contains the LSCV statistic for each density estimate, and can be used to produce an LSCV plot.

mcyc Usage lscvplot(..., alpha) Arguments ... alpha arguments to the lscv, locfit functions.

45

Matrix of smoothing parameters. The lscvplot function loops through calls to lscv, using each row of alpha as the smoothing parameter in turn. If alpha is provided as a vector, it will be converted to a one-column matrix, thus interpreting each component as a nearest neighbor smoothing parameter.

Value An object with class "gcvplot", containing the smoothing parameters and LSCV scores. The actual plot is produced using plot.gcvplot. See Also locfit, locfit.raw, gcv, lscv, plot.gcvplot

mcyc

Acc(De?)celeration of a Motorcycle Hitting a Wall

Description Measurements of the acceleration of a motorcycle as it hits a wall. Actually, rumored to be a concatenation of several such datasets. Usage data(mcyc) Format Data frame with time and accel variables. Source H\"ardle (1990). References H\"ardle, W. (1990). Applied Nonparametric Regression. Cambridge University Press.

46

mmsamp

mine

Fracture Counts in Coal Mines

Description The number of fractures in the upper seam of coal mines, and four predictor variables. This dataset can be modeled using Poisson regression. Usage data(mine) Format A dataframe with the response frac, and predictor variables extrp, time, seamh and inb. Source Myers (1990). References Myers, R. H. (1990). Classical and Modern Regression with Applications (Second edition). PWSKent Publishing, Boston.

mmsamp

Test dataset for minimax Local Regression

Description 50 observations, as used in Figure 13.1 of Loader (1999). Usage data(cltest) Format Data Frame with x and y variables. References Loader, C. (1999). Local Regression and Likelihood. Springer, New York.

morths

47

morths

Henderson and Sheppard Mortality Dataset

Description Observed mortality for 55 to 99. Usage data(morths) Format Data frame with age, n and number of deaths. Source Henderson and Sheppard (1919). References Henderson, R. and Sheppard, H. N. (1919). Graduation of mortality and other tables. Actuarial Society of America, New York.

none

Loct Evaluation Structure

Description none() is an evaluation structure for locfit.raw(), specifying no evaluation points. Only the initial parametric t is computed - this is the easiest and most efcient way to coerce Loct into producing a parametric regression t. Usage none() Examples
data(ethanol, package="locfit") # fit a fourth degree polynomial using locfit fit <- locfit(NOx~E,data=ethanol,deg=4,ev=none()) plot(fit,get.data=TRUE)

48

plot.eval

penny

Penny Thickness Dataset

Description For each year, 1945 to 1989, the thickness of two U.S. pennies was recorded. Usage data(penny) Format A dataframe. Source Scott (1992). References Scott (1992). Multivariate Density Estimation. Wiley, New York.

plot.eval

Plot evaluation points from a 2-d loct object.

Description This function is used to plot the evaluation structure generated by Loct for a two dimensional t. Vertices of the tree structure are displayed as O; pseudo-vertices as *. Usage plot.eval(x, add=FALSE, text=FALSE, ...) Arguments x add text ... See Also locfit. "locfit" object. If TRUE, add to existing plot. If TRUE, numbers will be added indicating the order points were added. Arguments passed to and from other methods.

plot.gcvplot Examples
data(ethanol, package="locfit") fit <- locfit(NOx ~ E + C, data=ethanol, scale= ) plot.eval(fit)

49

plot.gcvplot

Produce a cross-validation plot.

Description Plots the value of the GCV (or other statistic) in a gcvplot object against the degrees of freedom of the t.

Usage ## S3 method for class gcvplot plot(x, xlab = "Fitted DF", ylab = x$cri, ...)

Arguments x xlab ylab ... A gcvplot object, produced by gcvplot, aicplot etc. Text label for the x axis. Text label for the y axis. Other arguments to plot .

See Also locfit, locfit.raw, gcv, aicplot, cpplot, gcvplot, lcvplot

Examples
data(ethanol) plot(gcvplot(NOx~E,data=ethanol,alpha=seq( .2,1. ,by= . 5)))

50

plot.loct

plot.lfeval

Plot a Loct Evaluation Structure.

Description Plots the evaluation points from a locfit or lfeval structure, for one- or two-dimensional ts. Usage ## S3 method for class lfeval plot(x, add=FALSE, txt=FALSE, ...) Arguments x add txt ... Value "lfeval" object. See Also lfeval, locfit, print.lfeval A lfeval or locfit object If TRUE, the points will be added to the existing plot. Otherwise, a new plot is created. If TRUE, the points are annotated with numbers in the order they were entered into the t. Additional graphical parameters.

plot.locfit

Plot an object of class loct.

Description The plot.locfit function generates grids of ploting points, followed by a call to preplot.locfit. The returned object is then passed to plot.locfit.1d, plot.locfit.2d or plot.locfit.3d as appropriate. Usage ## S3 method for class locfit plot(x, xlim, pv, tv, m, mtv=6, band="none", tr=NULL, what = "coef", get.data=FALSE, f3d=(d == 2) && (length(tv) > ), ...)

plot.loct Arguments x xlim pv loct object.

51

Plotting limits. Eg. xlim=c( , ,1,1) plots over the unit square in two dimensions. Default is bounding box of the data. Panel variables, to be varied within each panel of a plot. May be specied as a character vector, or variable numbers. There must be one or two panel variables; default is all variables in one or two dimensions; Variable 1 in three or more dimensions. May by specied using either variable numbers or names. Trellis variables, to be varied from panel to panel of the plot. Controls the plot resolution (within panels, for trellis displays). Default is 100 points in one dimension; 40 points (per dimension) in two or more dimensions. Number of points for trellis variables; default 6. Type of condence bands to add to the plot. Default is "none". Other choices include "global" for bands using a global variance estimate; "local" for bands using a local variance estimate and "pred" for prediction bands (at present, using a global variance estimate). To obtain the global variance estimate for a t, use rv. This can be changed with rv<-. Condence bands, by default, are 95%, based on normal approximations and neglecting bias. To change the critical value or condence level, or to obtain simultaneous instead of pointwise condence, the critical value stored on the t must be changed. See the kappa and crit functions. Transformation function to use for plotting. Default is the inverse link function, or the identity function if derivatives are requested. What to plot. See predict.locfit. If TRUE, original data is added to the plot. Default: FALSE. Force the locfit.3d class on the prediction object, thereby generating a trellis style plot. Default: FALSE, unless a tv argument is provided. Not available in R. Other arguments to plot.locfit.1d, plot.locfit.2d or plot.locfit.3d as appropriate.

tv m mtv band

tr what get.data f3d

...

See Also locfit, plot.locfit.1d, plot.locfit.2d, plot.locfit.3d, lines.locfit, predict.locfit, preplot.locfit Examples
x <- rnorm(1 ) y <- dnorm(x) + rnorm(1 ) / 5 plot(locfit(y~x), band="global") x <- cbind(rnorm(1 ), rnorm(1 )) plot(locfit(~x), type="persp")

52

plot.loct.2d

plot.locfit.1d

Plot a one dimensional preplot.loct object.

Description This function is not usually called directly. It will be called automatically when plotting a onedimensional locfit or preplot.locfit object. Usage plot.locfit.1d(x, add=FALSE, main="", xlab="default", ylab=x$yname, type="l", ylim, lty=1, col=1, ...) Arguments x One dimensional preplot.locfit object. add If TRUE, the plot will be added to the existing plot. main, xlab, ylab, type, ylim, lty, col Graphical parameters passed on to plot (only if add=FALSE). ... See Also locfit, plot.locfit, preplot.locfit Additional graphical parameters to the plot function (only if add=FALSE).

plot.locfit.2d

Plot a two-dimensional "preplot.loct" object.

Description This function is not usually called directly. It will be called automatically when plotting onedimensional locfit or preplot.locfit objects. Usage plot.locfit.2d(x, type="contour", main, xlab, ylab, zlab=x$yname, ...) Arguments x type main xlab, ylab zlab ... Two dimensional preplot.locfit object. one of "contour", "persp", or "image". title for the plot. text labels for the x- and y-axes. if type="persp", the label for the z-axis. Additional arguments to the contour, persp or image functions.

plot.loct.3d See Also locfit, plot.locfit, preplot.locfit

53

plot.locfit.3d

Plot a high-dimensional "preplot.loct" object using trellis displays.

Description This function plots cross-sections of a Loct model (usually in three or more dimensions) using trellis displays. It is not usually called directly, but is invoked by plot.locfit. The R libraries lattice and grid provide a partial (at time of writing) implementation of trellis. Currently, this works with one panel variable. Usage plot.locfit.3d(x, main="", pv, tv, type = "level", pred.lab = x$vnames, resp.lab=x$yname, crit = 1.96, ...) Arguments x main pv tv type pred.lab resp.lab crit ... See Also plot.locfit, preplot.locfit "preplot.locfit" object. title for the plot. Panel variables. These are the variables (either one or two) that are varied within each panel of the display. Trellis variables. These are varied from panel to panel of the display. Type of display. When there are two panel variables, the choices are "contour", "level" and "persp". label for the predictor variable. label for the response variable. critical value for the condence level. graphical parameters passed to xyplot or contourplot.

54

plot.scb

plot.preplot.locfit

Plot a "preplot.loct" object.

Description The plot.locfit() function is implemented, roughly, as a call to preplot.locfit(), followed by a call to plot.locfitpred(). For most users, there will be little need to call plot.locfitpred() directly. Usage ## S3 method for class preplot.locfit plot(x, pv, tv, ...) Arguments x pv, tv, ... A preplot.locfit object, produced by preplot.locfit(). Other arguments to plot.locfit.1d, plot.locfit.2d or plot.locfit.3d as appropriate.

See Also locfit, plot.locfit, preplot.locfit, plot.locfit.1d, plot.locfit.2d, plot.locfit.3d.

plot.scb

Plot method for simultaneous condence bands

Description Plot method for simultaneous condence bands created by the scb function. Usage ## S3 method for class scb plot(x, add=FALSE, ...) Arguments x add ... See Also scb scb object created by scb. If TRUE, bands will be added to the existing plot. Arguments passed to and from other methods.

plotbyfactor Examples
# corrected confidence bands for a linear logistic model data(insect) fit <- scb(deaths ~ lconc, type=4, w=nins, data=insect, deg=1, family="binomial", kern="parm") plot(fit)

55

plotbyfactor

x-y scatterplot, colored by levels of a factor.

Description Produces a scatter plot of x-y data, with different classes given by a factor f. The different classes are identied by different colours and/or symbols. Usage plotbyfactor(x, y, f, data, col = 1:1 , pch = "O", add = FALSE, lg, xlab = deparse(substitute(x)), ylab = deparse(substitute(y)), log = "", ...) Arguments x y f data col pch add lg xlab, ylab log ... Examples
data(iris) plotbyfactor(petal.wid, petal.len, species, data=iris)

Variable for x axis. Variable for y axis. Factor (or variable for which as.factor() works). data frame for variables x, y, f. Default: sys.parent(). Color numbers to use in plot. Will be replicated if shorter than the number of levels of the factor f. Default: 1:10. Vector of plot characters. Replicated if necessary. Default: "O". If TRUE, add to existing plot. Otherwise, create new plot. Coordinates to place a legend. Default: Missing (no legend). Axes labels. Should the axes be in log scale? Use "x", "y", or "xy" to specify which axis to be in log scale. Other graphical parameters, labels, titles e.t.c.

56

predict.loct

points.locfit

Add loct points to existing plot

Description This function shows the points at which the local t was computed directly, rather than being interpolated. This can be useful if one is unsure of the validity of interpolation. Usage ## S3 method for class locfit points(x, tr, ...) Arguments x tr ... See Also locfit, plot.locfit, points "locfit" object. Should be a model with one predictor. Back transformation. Other arguments to the default points function.

predict.locfit

Prediction from a Loct object.

Description The locfit function computes a local t at a selected set of points (as dened by the ev argument). The predict.locfit function is used to interpolate from these points to any other points. The method is based on cubic hermite polynomial interpolation, using the estimates and local slopes at each t point. The motivation for this two-step procedure is computational speed. Depending on the sample size, dimension and tting procedure, the local tting method can be expensive, and it is desirable to keep the number of points at which the direct t is computed to a minimum. The interpolation method used by predict.locfit() is usually much faster, and can be computed at larger numbers of points. Usage ## S3 method for class locfit predict(object, newdata=NULL, where = "fitp", se.fit=FALSE, band="none", what="coef", ...)

preplot.loct Arguments object newdata Fitted object from locfit().

57

Points to predict at. Can be given in several forms: vector/matrix; list, data frame.

se.fit If TRUE, standard errors are computed along with the tted values. where, what, band arguments passed on to preplot.locfit. ... Value If se.fit=F, a numeric vector of predictors. If se.fit=T, a list with components fit, se.fit and residual.scale. Examples
data(ethanol, package="locfit") fit <- locfit(NOx ~ E, data=ethanol) predict(fit,c( .6, .8,1. ))

Additional arguments to preplot.locfit.

preplot.locfit

Prediction from a Loct object.

Description preplot.locfit can be called directly, although it is more usual to call plot.locfit or predict.locfit. The advantage of preplot.locfit is in S-Plus 5, where arithmetic and transformations can be performed on the "preplot.locfit" object. plot(preplot(fit)) is essentially synonymous with plot(fit). Usage ## S3 method for class locfit preplot(object, newdata=NULL, where, tr=NULL, what="coef", band="none", get.data=FALSE, f3d=FALSE, ...) Arguments object newdata where Fitted object from locfit(). Points to predict at. Can be given in several forms: vector/matrix; list, data frame. An alternative to newdata. Choices include "grid" for the grid lfmarg(object); "data" for the original data points and "fitp" for the direct tting points (ie. no interpolation). Transformation for likelihood models. Default is the inverse of the link function.

tr

58 what

preplot.loct.raw What to compute predicted values of. The default, what="coef", works with the tted curve itself. Other choices include "nlx" for the length of the weight diagram; "infl" for the inuence function; "band" for the bandwidth; "degr" for the local polynomial degree; "lik" for the maximized local likelihood; "rdf" for the local residual degrees of freedom and "vari" for the variance function. The interpolation algorithm for some of these quantities is questionable. Compute standard errors for the t and include condence bands on the returned object. Default is "none". Other choices include "global" for bands using a global variance estimate; "local" for bands using a local variance estimate and "pred" for prediction bands (at present, using a global variance estimate). To obtain the global variance estimate for a t, use rv. This can be changed with rv<-. Condence bands, by default, are 95%, based on normal approximations and neglecting bias. To change the critical value or condence level, or to obtain simultaneous instead of pointwise condence, the critical value stored on the t must be changed. See the kappa and crit functions. If TRUE, the original data is attached to the returned object, and added to the plot. If TRUE, sets a ag that forces ploting using the trellis style. Not available in R. arguments passed to and from other methods.

band

get.data f3d ... Value

An object with class "preplot.locfit", containing the predicted values and additional information used to construct the plot. See Also locfit, predict.locfit, plot.locfit.

preplot.locfit.raw

Prediction from a Loct object.

Description preplot.locfit.raw is an internal function used by predict.locfit and preplot.locfit. It should not normally be called directly. Usage ## S3 method for class locfit.raw preplot(object, newdata, where, what, band, ...)

print.gcvplot Arguments object newdata where what band ... Value A list containing raw output from the internal prediction routines. See Also locfit, predict.locfit, preplot.locfit. Fitted object from locfit(). New data points. Type of data provided in newdata. What to compute predicted values of.

59

Compute standard errors for the t and include condence bands on the returned object. Arguments passed to and from other methods.

print.gcvplot

Print method for gcvplot objects

Description Print method for "gcvplot" objects. Actually, equivalent to plot.gcvplot(). scb function. Usage ## S3 method for class gcvplot print(x, ...) Arguments x ... See Also gcvplot, plot.gcvplot summary.gcvplot gcvplot object. Arguments passed to and from other methods.

60

print.loct

print.lfeval

Print the Loct Evaluation Points.

Description Prints a matrix of the evaluation points from a locfit or lfeval structure. Usage ## S3 method for class lfeval print(x, ...) Arguments x ... Value Matrix of the t points. See Also lfeval, locfit, plot.lfeval A lfeval or locfit object Arguments passed to and from other methods.

print.locfit

Print method for "loct" object.

Description Prints a short summary of a "locfit" object. Usage ## S3 method for class locfit print(x, ...) Arguments x ... See Also locfit locfit object. Arguments passed to and from other methods.

print.preplot.loct

61

print.preplot.locfit

Print method for preplot.loct objects.

Description Print method for objects created by the preplot.locfit function. Usage ## S3 method for class preplot.locfit print(x, ...) Arguments x ... See Also preplot.locfit, predict.locfit "preplot.locfit" object. Arguments passed to and from other methods.

print.scb

Print method for simultaneous condence bands

Description Print method for simultaneous condence bands created by the scb function. Usage ## S3 method for class scb print(x, ...) Arguments x ... See Also scb "scb" object created by scb. Arguments passed to and from other methods.

62

rbox

print.summary.locfit

Print a Loct summary object.

Description Print method for "summary.locfit" objects. Usage ## S3 method for class summary.locfit print(x, ...) Arguments x ... See Also summary.locfit() Object from summary.locfit. Arguments passed to and from methods.

rbox

Local Regression, Likelihood and Density Estimation.

Description rbox() is used to specify a rectangular box evaluation structure for locfit.raw(). The structure begins by generating a bounding box for the data, then recursively divides the box to a desired precision. Usage rbox(cut= .8, type="tree", ll, ur) Arguments type If type="tree", the cells are recursively divided according to the bandwidths at each corner of the cell; see Chapter 11 of Loader (1999). If type="kdtree", the K-D tree structure used in Loess (Cleveland and Grosse, 1991) is used. Precision of the tree; a smaller value of cut results in a larger tree with more nodes being generated. Lower left corner of the initial cell. Length should be the number of dimensions of the data provided to locfit.raw(). Upper right corner of the initial cell. By default, ll and ur are generated as the bounding box for the data.

cut ll ur

regband References Loader, C. (1999). Local Regression and Likelihood. Springer, New York.

63

Cleveland, W. and Grosse, E. (1991). Computational Methods for Local Regression. Statistics and Computing 1. Examples
data(ethanol, package="locfit") plot.eval(locfit(NOx~E+C,data=ethanol,scale= ,ev=rbox(cut= .8))) plot.eval(locfit(NOx~E+C,data=ethanol,scale= ,ev=rbox(cut= .3)))

regband

Bandwidth selectors for local regression.

Description Function to compute local regression bandwidths for local linear regression, implemented as a front end to locfit(). This function is included for comparative purposes only. Plug-in selectors are based on awed logic, make unreasonable and restrictive assumptions and do not use the full power of the estimates available in Loct. Any relation between the results produced by this function and desirable estimates are entirely coincidental. Usage regband(formula, what = c("CP", "GCV", "GKK", "RSW"), deg=1, ...) Arguments formula what deg ... Value Vector of selected bandwidths. Model Formula (one predictor). Methods to use. Degree of t. Other Loct options.

64

right

residuals.locfit

Fitted values and residuals for a Loct object.

Description residuals.locfit is implemented as a front-end to fitted.locfit, with the type argument set. Usage ## S3 method for class locfit residuals(object, data=NULL, type="deviance", ...) Arguments object data type locfit object. The data frame for the original t. Usually, shouldnt be needed. Type of t or residuals to compute. The default is "fit" for fitted.locfit, and "dev" for residuals.locfit. Other choices include "pear" for Pearson residuals; "raw" for raw residuals, "ldot" for likelihood derivative; "d2" for the deviance residual squared; lddot for the likelihood second derivative. Generally, type should only be used when what="coef". arguments passed to and from other methods.

... Value

A numeric vector of the residuals.

right

One-sided right smooth for a Loct model.

Description The right() function is used in a loct model formula to specify a one-sided smooth: when tting at a point x, only data points with xi x should be used. This can be useful in estimating points of discontinuity, and in cross-validation for forecasting a time series. right(x) is equivalent to lp(x,style="right"). When using this function, it will usually be necessary to specify an evaluation structure, since the t is not smooth and locts interpolation methods are unreliable. Also, it is usually best to use deg= or deg=1, otherwise the ts may be too variable. If nearest neighbor bandwidth specication is used, it does not recognize right(). Usage right(x,...)

rv Arguments x ... See Also lfbas, locfit, left Examples

# compute left and right smooths data(penny) xev <- (1945:1988)+ .5 fitl <- locfit(thickness~left(year,h=1 ,deg=1), ev=xev, data=penny) fitr <- locfit(thickness~right(year,h=1 ,deg=1),ev=xev, data=penny) # plot the squared difference, to show the change points. plot( xev, (predict(fitr,where="ev") - predict(fitl,where="ev"))^2 )

65

numeric variable. Other arguments to lp().

rv

Residual variance from a loct object.

Description As part of the locfit tting procedure, an estimate of the residual variance is computed; the rv function extracts the variance from the "locfit" object. The estimate used is the residual sum of squares (or residual deviance, for quasi-likelihood models), divided by the residual degrees of freedom. For likelihood (not quasi-likelihood) models, the estimate is 1.0. Usage rv(fit) Arguments fit Value Returns the residual variance estimate from the "locfit" object. See Also loct, rv<"locfit" object.

66 Examples
data(ethanol) fit <- locfit(NOx~E,data=ethanol) rv(fit)

scb

rva

Substitute variance estimate on a loct object.

Description By default, Loct uses the normalized residual sum of squares as the variance estimate when constructing condence intervals. In some cases, the user may like to use alternative variance estimates; this function allows the default value to be changed. Usage rv(fit) <- 1.2345 Arguments fit See Also loct(), rv(), plot.loct() "locfit" object.

scb

Simultaneous Condence Bands

Description scb is implemented as a front-end to locfit, to compute simultaneous condence bands using the tube formula method and extensions, based on Sun and Loader (1994). Usage scb(x, ..., ev = lfgrid(2 ), simul = TRUE, type = 1)

sjpi Arguments x ... ev simul type

67

A numeric vector or matrix of predictors (as in locfit.raw), or a model formula (as in locfit). Additional arguments to locfit.raw. The evaluation structure to use. See locfit.raw. Should the coverage be simultaneous or pointwise? Type of condence bands. type= computes pointwise 95% bands. type=1 computes basic simultaneous bands with no corrections. type=2,3,4 are the centered and corrected bands for parametric regression models listed in Table 3 of Sun, Loader and McCormick (2000).

Value A list containing the evaluation points, t, standard deviations and upper and lower condence bounds. The class is "scb"; methods for printing and ploting are provided. References Sun J. and Loader, C. (1994). Simultaneous condence bands in linear regression and smoothing. The Annals of Statistics 22, 1328-1345. Sun, J., Loader, C. and McCormick, W. (2000). Condence bands in generalized linear models. The Annals of Statistics 28, 429-460. See Also locfit, print.scb, plot.scb. Examples
# corrected confidence bands for a linear logistic model data(insect) fit <- scb(deaths~lp(lconc,deg=1), type=4, w=nins, data=insect,family="binomial",kern="parm") plot(fit)

sjpi

Sheather-Jones Plug-in bandwidth criterion.

Description Given a dataset and set of pilot bandwidths, this function computes a bandwidth via the plug-in method, and the assumed pilot relationship of Sheather and Jones (1991). The S-J method chooses the bandwidth at which the two intersect. The purpose of this function is to demonstrate the sensitivity of plug-in methods to pilot bandwidths and assumptions. This function does not provide a reliable method of bandwidth selection.

68 Usage sjpi(x, a) Arguments x a Value data vector vector of pilot bandwidths

smooth.lf

A matrix with four columns; the number of rows equals the length of a. The rst column is the plug-in selected bandwidth. The second column is the pilot bandwidths a. The third column is the pilot bandwidth according to the assumed relationship of Sheather and Jones. The fourth column is an intermediate calculation. References Sheather, S. J. and Jones, M. C. (1991). A reliable data-based bandwidth selection method for kernel density estimation. JRSS-B 53, 683-690. See Also locfit, locfit.raw, lcvplot Examples
# Fig 1 .2 (S-J parts) from Loader (1999). data(geyser) gf <- 2.5 a <- seq( . 5, .7, length=1 ) z <- sjpi(geyser, a) # the plug-in curve. Multiplying by gf=2.5 corresponds to Locfits standard # scaling for the Gaussian kernel. plot(gf*z[, 2], gf*z[, 1], type = "l", xlab = "Pilot Bandwidth k", ylab = "Bandwidth h") # Add the assumed curve. lines(gf * z[, 3], gf * z[, 1], lty = 2) legend(gf* . 5, gf* .4, lty = 1:2, legend = c("Plug-in", "SJ assumed"))

smooth.lf

Local Regression, Likelihood and Density Estimation.

Description smooth.lf is a simple interface to the Loct library. The input consists of a predictor vector (or matrix) and response. The output is a list with vectors of tting points and tted values. Most locfit.raw options are valid.

spence.15 Usage smooth.lf(x, y, xev=x, direct=FALSE, ...) Arguments x y xev direct ... Value Vector (or matrix) of the independent variable(s).

69

Response variable. If omitted, x is treated as the response and the predictor variable is 1:n. Fitting Points. Default is the data vector x. Logical variable. If T, local regression is performed directly at each tting point. If F, the standard Loct method combining tting and interpolation is used. Other arguments to locfit.raw().

A list with components x (tting points) and y (tted values). Also has a call component, so update() will work. See Also locfit(), locfit.raw(), density.lf(). Examples
# using smooth.lf() to fit a local likelihood model. data(morths) fit <- smooth.lf(morths$age, morths$deaths, weights=morths$n, family="binomial") plot(fit,type="l") # update with the direct fit fit1 <- update(fit, direct=TRUE) lines(fit1,col=2) print(max(abs(fit$y-fit1$y)))

spence.15

Spencers 15 point graduation rule.

Description Spencers 15 point rule is a weighted moving average operation for a sequence of observations equally spaced in time. The average at time t depends on the observations at times t-7,...,t+7. Except for boundary effects, the function will reproduce polynomials up to degree 3. Usage spence.15(y)

70 Arguments y Value Data vector of observations at equally spaced points.

spence.21

A vector with the same length as the input vector, representing the graduated (smoothed) values. References Spencer, J. (1904). On the graduation of rates of sickness and mortality. Journal of the Institute of Actuaries 38, 334-343. See Also spence.21, spencer, Examples
data(spencer) yy <- spence.15(spencer$mortality) plot(spencer$age, spencer$mortality) lines(spencer$age, yy)

spence.21

Spencers 21 point graduation rule.

Description Spencers 21 point rule is a weighted moving average operation for a sequence of observations equally spaced in time. The average at time t depends on the observations at times t-11,...,t+11. Except for boundary effects, the function will reproduce polynomials up to degree 3. Usage spence.21(y) Arguments y Value A vector with the same length as the input vector, representing the graduated (smoothed) values. References Spencer, J. (1904). On the graduation of rates of sickness and mortality. Journal of the Institute of Actuaries 38, 334-343. Data vector of observations at equally spaced points.

spencer See Also spence.15, spencer, Examples

data(spencer) yy <- spence.21(spencer$mortality) plot(spencer$age, spencer$mortality) lines(spencer$age, yy)

71

spencer

Spencers Mortality Dataset

Description Observed mortality rates for ages 20 to 45. Usage data(spencer) Format Data frame with age and mortality variables. Source Spencer (1904). References Spencer, J. (1904). On the graduation of rates of sickness and mortality. Journal of the Institute of Actuaries 38, 334-343.

stamp

Stamp Thickness Dataset

Description Thicknesses of 482 postage stamps of the 1872 Hidalgo issue of Mexico. Usage data(stamp)

72 Format

summary.gcvplot

Data frame with thick (stamp thickness) and count (number of stamps) variables. Source Izenman and Sommer (1988). References Izenman, A. J. and Sommer, C. J. (1988). Philatelic mixtures and multimodal densities. Journal of the American Statistical Association 73, 602-606.

store

Save S functions.

Description Ive gotta keep track of this mess somehow! Usage store(data=FALSE, grand=FALSE) Arguments data grand whether data objects are to be saved. whether everything is to be saved.

summary.gcvplot

Summary method for a gcvplot structure.

Description Computes a short summary for a generalized cross-validation plot structure Usage ## S3 method for class gcvplot summary(object, ...) Arguments object ... A gcvplot structure produced by a call to gcvplot, cpplot e.t.c. arugments to and from other methods.

summary.loct Value

73

A matrix with two columns; one row for each t computed in the gcvplot call. The rst column is the tted degrees of freedom; the second is the GCV or other criterion computed.

See Also locfit, gcv, gcvplot

Examples
data(ethanol) summary(gcvplot(NOx~E,data=ethanol,alpha=seq( .2,1. ,by= . 5)))

summary.locfit

Print method for a loct object.

Description Prints a short summary of a "locfit" object.

Usage ## S3 method for class locfit summary(object, ...)

Arguments object ... locfit object. arguments passed to and from methods.

Value A summary.locfit object, containg a short summary of the locfit object.

74

trimod

summary.preplot.locfit Summary method for a preplot.loct object.

Description Prints a short summary of a "preplot.locfit" object. Usage ## S3 method for class preplot.locfit summary(object, ...) Arguments object ... Value The tted values from a preplot.locfit object. preplot.locfit object. arguments passed to and from methods.

trimod

Generated sample from a bivariate trimodal normal mixture

Description This is a random sample from a mixture of three bivariate standard normal components; the sample was used for the examples in Loader (1996). Format Data frame with 225 observations and variables x0, x1. Source Randomly generated in S. References Loader, C. R. (1996). Local Likelihood Density Estimation. Annals of Statistics 24, 1602-1618.

xbar

75

xbar

Loct Evaluation Structure

Description xbar() is an evaluation structure for locfit.raw(), evaluating the t at a single point, namely, the average of each predictor variable. Usage xbar()

Index
Topic datasets ais, 5 bad, 7 border, 7 chemdiab, 8 claw54, 8 cldem, 9 cltest, 9 cltrain, 10 co2, 10 diab, 15 ethanol, 16 gam.slist, 19 geyser, 20 geyser.round, 21 heart, 22 insect, 23 iris, 23 kangaroo, 24 livmet, 34 mcyc, 45 mine, 46 mmsamp, 46 morths, 47 penny, 48 spencer, 71 stamp, 71 trimod, 74 Topic htest aic, 4 aicplot, 4 cp, 11 cpplot, 12 gcv, 19 gcvplot, 20 kdeb, 26 lcv, 28 lcvplot, 28 lscv, 43 76 lscv.exact, 44 lscvplot, 44 regband, 63 sjpi, 67 Topic math expit, 16 Topic methods plot.gcvplot, 49 plot.locfit, 50 plot.locfit.1d, 52 plot.locfit.2d, 52 plot.locfit.3d, 53 plot.scb, 54 preplot.locfit.raw, 58 print.gcvplot, 59 print.locfit, 60 print.preplot.locfit, 61 print.scb, 61 print.summary.locfit, 62 summary.gcvplot, 72 summary.locfit, 73 summary.preplot.locfit, 74 Topic models ang, 6 cpar, 11 formula.locfit, 18 gam.lf, 18 left, 29 lf, 30 locfit.matrix, 37 lp, 42 right, 64 Topic smooth crit, 13 dat, 14 density.lf, 14 fitted.locfit, 17 hatmatrix, 22 kappa , 25

INDEX km.mrl, 27 lfeval, 30 lfgrid, 31 lfknots, 32 lflim, 32 lfmarg, 33 lines.locfit, 33 locfit, 35 locfit.censor, 36 locfit.quasi, 38 locfit.raw, 39 locfit.robust, 41 none, 47 plot.eval, 48 plot.lfeval, 50 plot.preplot.locfit, 54 plotbyfactor, 55 points.locfit, 56 predict.locfit, 56 preplot.locfit, 57 print.lfeval, 60 rbox, 62 residuals.locfit, 64 rv, 65 rva, 66 scb, 66 smooth.lf, 68 spence.15, 69 spence.21, 70 store, 72 xbar, 75 aic, 4, 4, 5, 12 aicplot, 4, 4, 49 ais, 5 ang, 6, 42 bad, 7 border, 7 chemdiab, 8 claw54, 8 cldem, 9 cltest, 9 cltrain, 10 co2, 10 cp, 11, 12 cpar, 11 cpplot, 11, 12, 49, 72 crit, 13, 13, 25, 51, 58 crit<- (crit), 13 crit<-, 13, 25 dat, 14, 40 density.lf, 14, 69 diab, 15 ethanol, 16 expit, 16 fitted.locfit, 17, 37, 38, 64 formula.locfit, 18 gam.lf, 18, 30 gam.slist, 19 gcv, 5, 12, 19, 20, 29, 45, 49, 73 gcvplot, 19, 20, 49, 59, 72, 73 geyser, 20, 21 geyser.round, 21 hatmatrix, 22 heart, 22 insect, 23 iris, 23 kangaroo, 24 kappa , 13, 25, 51, 58 kdeb, 26 km.mrl, 27, 37

77

lcv, 28, 28, 29 lcvplot, 28, 28, 49, 68 left, 29, 42, 65 lf, 18, 19, 30 lfbas, 40, 65 lfeval, 30, 50, 60 lfgrid, 15, 31, 40 lfknots, 32 lflim, 32 lfmarg, 33, 57 lines, 34 lines.locfit, 22, 33, 51 livmet, 34 llines.locfit (lines.locfit), 33 locfit, 46, 1113, 15, 1820, 22, 25, 2834, 35, 3639, 4143, 45, 4854, 5660, 63, 6569, 73 locfit.censor, 27, 36

78 locfit.matrix, 37 locfit.quasi, 38 locfit.raw, 4, 5, 11, 12, 14, 15, 1820, 22, 2831, 3538, 39, 4143, 45, 47, 49, 62, 6769, 75 locfit.robust, 41 lp, 6, 29, 40, 41, 42, 44, 65 lscv, 43, 44, 45 lscv.exact, 43, 44 lscvplot, 43, 44, 44 mcyc, 45 mine, 46 mmsamp, 46 morths, 47 none, 40, 47 penny, 48 plot, 49, 52 plot.eval, 48 plot.gcvplot, 5, 12, 20, 29, 45, 49, 59 plot.lfeval, 31, 50, 60 plot.locfit, 13, 25, 33, 34, 50, 5254, 5658, 66 plot.locfit.1d, 22, 50, 51, 52, 54 plot.locfit.2d, 22, 50, 51, 52, 54 plot.locfit.3d, 22, 50, 51, 53, 54 plot.preplot.locfit, 54 plot.scb, 54, 67 plotbyfactor, 55 points, 56 points.locfit, 56 predict.locfit, 22, 51, 56, 5759, 61 preplot.locfit, 5054, 57, 5759, 61 preplot.locfit.raw, 58 print.gcvplot, 59 print.lfeval, 31, 50, 60 print.locfit, 60 print.preplot.locfit, 61 print.scb, 61, 67 print.summary.locfit, 62 rbox, 40, 62 regband, 63 residuals.locfit, 17, 38, 64, 64 right, 29, 64 rv, 40, 51, 58, 65, 66 rv<-, 65 rv<- (rva), 66 rv<-, 51, 58 rva, 66 scb, 54, 59, 61, 66 sjpi, 67 smooth.lf, 40, 68 spence.15, 69, 71 spence.21, 70, 70 spencer, 70, 71, 71 stamp, 71 store, 72 summary.gcvplot, 20, 59, 72 summary.locfit, 62, 73 summary.preplot.locfit, 74 trimod, 74 xbar, 75

INDEX