1\name{plot.efp} 2\alias{plot.efp} 3\alias{lines.efp} 4\encoding{latin1} 5\title{Plot Empirical Fluctuation Process} 6\description{Plot and lines method for objects of class \code{"efp"}} 7\usage{ 8\method{plot}{efp}(x, alpha = 0.05, alt.boundary = FALSE, boundary = TRUE, 9 functional = "max", main = NULL, ylim = NULL, 10 ylab = "Empirical fluctuation process", ...) 11\method{lines}{efp}(x, functional = "max", ...) 12} 13 14\arguments{ 15 \item{x}{an object of class \code{"efp"}.} 16 \item{alpha}{numeric from interval (0,1) indicating the confidence level for 17 which the boundary of the corresponding test will be computed.} 18 \item{alt.boundary}{logical. If set to \code{TRUE} alternative boundaries 19 (instead of the standard linear boundaries) will be plotted (for CUSUM 20 processes only).} 21 \item{boundary}{logical. If set to \code{FALSE} the boundary will be computed 22 but not plotted.} 23 \item{functional}{indicates which functional should be applied to the 24 process before plotting and which boundaries should be used. If set to \code{NULL} 25 a multiple process with boundaries for the \code{"max"} functional is plotted. 26 For more details see below.} 27 \item{main, ylim, ylab, ...}{high-level \code{\link{plot}} function 28 parameters.} 29} 30 31\details{Plots are available for the \code{"max"} functional for all process types. 32For Brownian bridge type processes the maximum or mean squared Euclidean norm 33(\code{"maxL2"} and \code{"meanL2"}) can be used for aggregating before plotting. 34No plots are available for the \code{"range"} functional. 35 36Alternative boundaries that are proportional to the standard deviation 37of the corresponding limiting process are available for processes with Brownian 38motion or Brownian bridge limiting processes. 39} 40 41\value{\code{\link{efp}} returns an object of class \code{"efp"} which inherits 42from the class \code{"ts"} or \code{"mts"} respectively. The function 43\code{\link{plot}} has a method to plot the 44empirical fluctuation process; with \code{sctest} the corresponding test for 45structural change can be performed.} 46 47\references{Brown R.L., Durbin J., Evans J.M. (1975), Techniques for 48testing constancy of regression relationships over time, \emph{Journal of the 49Royal Statistical Society}, B, \bold{37}, 149-163. 50 51Chu C.-S., Hornik K., Kuan C.-M. (1995), MOSUM tests for parameter 52constancy, \emph{Biometrika}, \bold{82}, 603-617. 53 54Chu C.-S., Hornik K., Kuan C.-M. (1995), The moving-estimates test for 55parameter stability, \emph{Econometric Theory}, \bold{11}, 669-720. 56 57Kr�mer W., Ploberger W., Alt R. (1988), Testing for structural change in 58dynamic models, \emph{Econometrica}, \bold{56}, 1355-1369. 59 60Kuan C.-M., Hornik K. (1995), The generalized fluctuation test: A 61unifying view, \emph{Econometric Reviews}, \bold{14}, 135 - 161. 62 63Kuan C.-M., Chen (1994), Implementing the fluctuation and moving estimates 64tests in dynamic econometric models, \emph{Economics Letters}, \bold{44}, 65235-239. 66 67Ploberger W., Kr�mer W. (1992), The CUSUM test with OLS residuals, 68\emph{Econometrica}, \bold{60}, 271-285. 69 70Zeileis A., Leisch F., Hornik K., Kleiber C. (2002), \code{strucchange}: 71An R Package for Testing for Structural Change in Linear Regression Models, 72\emph{Journal of Statistical Software}, \bold{7}(2), 1-38. 73URL \url{http://www.jstatsoft.org/v07/i02/}. 74 75Zeileis A. (2004), Alternative Boundaries for CUSUM Tests, 76\emph{Statistical Papers}, \bold{45}, 123--131. 77} 78 79\seealso{\code{\link{efp}}, \code{\link{boundary.efp}}, 80\code{\link{sctest.efp}}} 81 82\examples{ 83## Load dataset "nhtemp" with average yearly temperatures in New Haven 84data("nhtemp") 85## plot the data 86plot(nhtemp) 87 88## test the model null hypothesis that the average temperature remains 89## constant over the years 90## compute Rec-CUSUM fluctuation process 91temp.cus <- efp(nhtemp ~ 1) 92## plot the process 93plot(temp.cus, alpha = 0.01) 94## and calculate the test statistic 95sctest(temp.cus) 96 97## compute (recursive estimates) fluctuation process 98## with an additional linear trend regressor 99lin.trend <- 1:60 100temp.me <- efp(nhtemp ~ lin.trend, type = "fluctuation") 101## plot the bivariate process 102plot(temp.me, functional = NULL) 103## and perform the corresponding test 104sctest(temp.me) 105} 106\keyword{hplot} 107