Nothing
R/lpdensity_methods.R
In lpdensity: Local Polynomial Density Estimation and Inference
Defines functions
confint.lpdensity
vcov.lpdensity
coef.lpdensity
plot.lpdensity
summary.lpdensity
print.lpdensity
Documented in
coef.lpdensity
confint.lpdensity
plot.lpdensity
print.lpdensity
summary.lpdensity
vcov.lpdensity
################################################################################
#' Print Method for Local Polynomial Density Estimation and Inference
#'
#' @description The print method for local polynomial density objects.
#'
#' @param x Class "lpdensity" object, obtained from calling \code{\link{lpdensity}}.
#' @param ... Additional options.
#'
#' @author
#' Matias D. Cattaneo, Princeton University. \email{cattaneo@princeton.edu}.
#'
#' Michael Jansson, University of California Berkeley. \email{mjansson@econ.berkeley.edu}.
#'
#' Xinwei Ma (maintainer), University of California San Diego. \email{x1ma@ucsd.edu}.
#'
#' @seealso \code{\link{lpdensity}} for local polynomial density estimation.
#'
#' Supported methods: \code{\link{coef.lpdensity}}, \code{\link{confint.lpdensity}},
#' \code{\link{plot.lpdensity}}, \code{\link{print.lpdensity}}, \code{\link{summary.lpdensity}},
#' \code{\link{vcov.lpdensity}}.
#'
#' @examples
#' # Generate a random sample
#' set.seed(42); X <- rnorm(2000)
#'
#' # Estimate density and report results
#' print(lpdensity(data = X, bwselect = "imse-dpi"))
#'
#' @export
print.lpdensity <- function(x, ...) {
cat("Call: lpdensity\n\n")
cat(paste("Sample size ", x$opt$n, "\n", sep=""))
cat(paste("Polynomial order for point estimation (p=) ", x$opt$p, "\n", sep=""))
cat(paste("Order of derivative estimated (v=) ", x$opt$v, "\n", sep=""))
if (x$opt$q > 0) {
cat(paste("Polynomial order for confidence interval (q=) ", x$opt$q, "\n", sep=""))
} else {
cat(paste("Polynomial order for confidence interval (q=) ", "p", "\n", sep=""))
}
cat(paste("Kernel function ", x$opt$kernel, "\n", sep=""))
if (x$opt$scale != 1) {
cat(paste("Scaling factor ", x$opt$scale, "\n", sep=""))
}
cat(paste("Bandwidth method ", x$opt$bwselect, "\n", sep=""))
cat("\n")
cat("Use summary(...) to show estimates.\n")
}
################################################################################
#' Summary Method for Local Polynomial Density Estimation and Inference
#'
#' @description The summary method for local polynomial density objects.
#'
#' @param object Class "lpdensity" object, obtained from calling \code{\link{lpdensity}}.
#' @param ... Additional options, including (i) \code{grid} specifies a subset of grid points
#' to display results; (ii) \code{gridIndex} specifies the indices of grid points
#' to display results; (iii) \code{alpha} specifies the significance level; (iv)
#' \code{CIuniform} specifies whether displaying pointwise confidence intervals (\code{FALSE}, default) or
#' the uniform confidence band (\code{TRUE}); (v) \code{CIsimul} specifies the number of simulations used
#' to construct critical values (default is \code{2000}).
#'
#' @author
#' Matias D. Cattaneo, Princeton University. \email{cattaneo@princeton.edu}.
#'
#' Michael Jansson, University of California Berkeley. \email{mjansson@econ.berkeley.edu}.
#'
#' Xinwei Ma (maintainer), University of California San Diego. \email{x1ma@ucsd.edu}.
#'
#' @seealso \code{\link{lpdensity}} for local polynomial density estimation.
#'
#' Supported methods: \code{\link{coef.lpdensity}}, \code{\link{confint.lpdensity}},
#' \code{\link{plot.lpdensity}}, \code{\link{print.lpdensity}}, \code{\link{summary.lpdensity}},
#' \code{\link{vcov.lpdensity}}.
#'
#' @examples
#' # Generate a random sample
#' set.seed(42); X <- rnorm(2000)
#'
#' # Estimate density and report results
#' est1 <- lpdensity(data = X, bwselect = "imse-dpi")
#' summary(est1)
#'
#' # Report results for a subset of grid points
#' summary(est1, grid=est1$Estimate[4:10, "grid"])
#' summary(est1, gridIndex=4:10)
#'
#' # Report the 99% uniform confidence band
#' set.seed(42) # fix the seed for simulating critical values
#' summary(est1, alpha=0.01, CIuniform=TRUE)
#'
#' @export
summary.lpdensity <- function(object, ...) {
x <- object
args <- list(...)
if (is.null(args[['alpha']])) { alpha <- 0.05 } else { alpha <- args[['alpha']] }
if (is.null(args[['sep']])) { sep <- 5 } else { sep <- args[['sep']] }
if (is.null(args[['CIuniform']])) { CIuniform <- FALSE } else { CIuniform <- args[['CIuniform']] }
if (is.null(args[['CIsimul']])) { CIsimul <- 2000 } else { sep <- args[['CIsimul']] }
if (is.null(args[['grid']]) & is.null(args[['gridIndex']])) {
gridIndex <- 1:nrow(x$Estimate)
} else if (is.null(args[['grid']]) & !is.null(args[['gridIndex']])) {
gridIndex <- args[['gridIndex']]
if (is.null(gridIndex)) {
gridIndex <- 1:nrow(x$Estimate)
} else if (!all(gridIndex %in% 1:nrow(x$Estimate))) {
stop(paste("Option gridIndex incorrectly specified. Should be integers between 1 and ", nrow(x$Estimate), ".\n", sep=""))
}
} else {
grid <- args[['grid']]
if (is.null(grid)) {
gridIndex <- 1:nrow(x$Estimate)
} else if (!is.numeric(grid)) {
stop("Option grid incorrectly specified.\n")
} else {
gridIndex <- rep(NA, length(grid))
if (min(grid) < min(x$Estimate[, "grid"]) | max(grid) > max(x$Estimate[, "grid"])) {
warning("The reporting range exceeds the original estimation range. Option summary(..., grid=) should be within the estimation range specified by lpdensity(..., grid=).\n")
}
for (j in 1:length(grid)) {
gridIndex[j] <- which.min(abs(x$Estimate[, "grid"]-grid[j]))
#gridIndex <- unique(gridIndex)
}
}
}
cat("Call: lpdensity\n\n")
cat(paste("Sample size (n=) ", x$opt$n, "\n", sep=""))
cat(paste("Polynomial order for point estimation (p=) ", x$opt$p, "\n", sep=""))
if (x$opt$v == 0) {
cat(paste("Distribution function estimated (v=) ", x$opt$v, "\n", sep=""))
} else if (x$opt$v == 1) {
cat(paste("Density function estimated (v=) ", x$opt$v, "\n", sep=""))
} else {
cat(paste("Order of derivative estimated (v=) ", x$opt$v, "\n", sep=""))
}
if (x$opt$q > x$opt$p) {
cat(paste("Polynomial order for confidence interval (q=) ", x$opt$q, "\n", sep=""))
} else {
cat(paste("Polynomial order for confidence interval (q=) ", "p", "\n", sep=""))
}
cat(paste("Kernel function ", x$opt$kernel, "\n", sep=""))
if (x$opt$scale != 1) {
cat(paste("Scaling factor ", x$opt$scale, "\n", sep=""))
}
cat(paste("Bandwidth selection method ", x$opt$bwselect, "\n", sep=""))
cat("\n")
### compute CI
if (CIuniform) {
if (length(CIsimul) == 0) { CIsimul <- 2000 }
if (!is.numeric(CIsimul) | is.na(CIsimul)) {
warning("Option CIsimul incorrectly specified. Will only plot pointwise confidence intervals.\n")
CIuniform <- FALSE
z_val <- qnorm(1 - alpha/2)
} else if (ceiling(CIsimul)<2) {
warning("Option CIsimul incorrectly specified. Will only plot pointwise confidence intervals.\n")
CIuniform <- FALSE
z_val <- qnorm(1 - alpha/2)
} else {
CIsimul <- ceiling(CIsimul)
if (x$opt$q == x$opt$p) {
x$CovMat_q <- x$CovMat_p
x$Estimate[, "se_q"] <- x$Estimate[, "se_p"]
}
corrMat <- sweep(sweep(x$CovMat_q, MARGIN=1, FUN="*", STATS=1/x$Estimate[, "se_q"]), MARGIN=2, FUN="*", STATS=1/x$Estimate[, "se_q"])
normalSimu <- try(
mvrnorm(n=CIsimul, mu=rep(0,nrow(corrMat)), Sigma=corrMat),
silent=TRUE)
if (is.character(normalSimu)) {
print(normalSimu)
warning("Variance-Covariance is not positive semidefinite. Will only plot pointwise confidence intervals.\n")
CIuniform <- FALSE
z_val <- qnorm(1 - alpha/2)
} else {
z_val <- quantile(apply(normalSimu, MARGIN=1, FUN=function(x) {max(abs(x))}), 1 - alpha)
}
}
} else {
z_val <- qnorm(1 - alpha/2)
}
if (x$opt$q == x$opt$p) {
CI_l <- x$Estimate[, "f_p"] - x$Estimate[, "se_p"] * z_val
CI_r <- x$Estimate[, "f_p"] + x$Estimate[, "se_p"] * z_val
} else {
CI_l <- x$Estimate[, "f_q"] - x$Estimate[, "se_q"] * z_val
CI_r <- x$Estimate[, "f_q"] + x$Estimate[, "se_q"] * z_val
}
flagNotAllLocalSampleSizeSame <- !all(x$Estimate[, "nh"] == x$Estimate[, "nhu"])
### print output
cat(paste(rep("=", 14 + 10 + 8 + 8*flagNotAllLocalSampleSizeSame + 10 + 10 + 25), collapse="")); cat("\n")
cat(format(" ", width= 14 ))
cat(format(" ", width= 10 ))
cat(format(" ", width= 8 + 8*flagNotAllLocalSampleSizeSame ))
cat(format("Point", width= 10, justify="right"))
cat(format("Std." , width= 10, justify="right"))
if (x$opt$q > x$opt$p) {
cat(format("Robust B.C."
, width=25, justify="centre"))
cat("\n")
} else {
cat(format("Conventional"
, width=25, justify="centre"))
cat("\n")
}
cat(format("Index Grid" , width=14, justify="right"))
cat(format("B.W." , width=10, justify="right"))
cat(format("Eff.n" , width=8 , justify="right"))
if (flagNotAllLocalSampleSizeSame) {
cat(format("Uniq.n" , width=8 , justify="right"))
}
cat(format("Est." , width=10, justify="right"))
cat(format("Error" , width=10, justify="right"))
if (CIuniform) {
cat(format(paste("[ Unif. ", floor((1-alpha)*100), "%", " C.I. ]", sep="")
, width=25, justify="centre"))
} else {
cat(format(paste("[ ", floor((1-alpha)*100), "%", " C.I. ]", sep="")
, width=25, justify="centre"))
}
cat("\n")
if (all(x$Estimate[, "nh"] == x$Estimate[, "nhu"])) {
cat(paste(rep("=", 14 + 10 + 8 + 10 + 10 + 25), collapse="")); cat("\n")
jj <- 1
for (j in gridIndex) {
cat(format(toString(j), width=4))
cat(format(sprintf("%6.4f", x$Estimate[j, "grid"]), width=10, justify="right"))
cat(format(sprintf("%6.4f", x$Estimate[j, "bw"]) , width=10, justify="right"))
cat(format(sprintf("%8.0f", x$Estimate[j, "nh"]) , width=8 , justify="right"))
cat(format(sprintf("%6.4f", x$Estimate[j, "f_p"]) , width=10, justify="right"))
cat(format(
paste(sprintf("%6.4f", x$Estimate[j, "se_p"]), sep=""), width=10, justify="right"))
cat(format(
paste(sprintf("%7.4f", CI_l[j]), " , ", sep="") , width=14, justify="right"))
cat(format(
paste(sprintf("%7.4f", CI_r[j]), sep=""), width=11, justify="left"))
cat("\n")
if (is.numeric(sep)) if (sep > 0) if (jj %% sep == 0) {
cat(paste(rep("-", 14 + 10 + 8 + 10 + 10 + 25), collapse="")); cat("\n")
}
jj <- jj + 1
}
cat(paste(rep("=", 14 + 10 + 8 + 10 + 10 + 25), collapse="")); cat("\n")
} else {
cat(paste(rep("=", 14 + 10 + 8 + 8 + 10 + 10 + 25), collapse="")); cat("\n")
jj <- 1
for (j in gridIndex) {
cat(format(toString(j), width=4))
cat(format(sprintf("%6.4f", x$Estimate[j, "grid"]), width=10, justify="right"))
cat(format(sprintf("%6.4f", x$Estimate[j, "bw"]) , width=10, justify="right"))
cat(format(sprintf("%8.0f", x$Estimate[j, "nh"]) , width=8 , justify="right"))
cat(format(sprintf("%8.0f", x$Estimate[j, "nhu"]) , width=8 , justify="right"))
cat(format(sprintf("%6.4f", x$Estimate[j, "f_p"]) , width=10, justify="right"))
cat(format(
paste(sprintf("%6.4f", x$Estimate[j, "se_p"]), sep=""), width=10, justify="right"))
cat(format(
paste(sprintf("%7.4f", CI_l[j]), " , ", sep="") , width=14, justify="right"))
cat(format(
paste(sprintf("%7.4f", CI_r[j]), sep=""), width=11, justify="left"))
cat("\n")
if (is.numeric(sep)) if (sep > 0) if (jj %% sep == 0) {
cat(paste(rep("-", 14 + 10 + 8 + 8 + 10 + 10 + 25), collapse="")); cat("\n")
}
jj <- jj + 1
}
cat(paste(rep("=", 14 + 10 + 8 + 8 + 10 + 10 + 25), collapse="")); cat("\n")
}
}
################################################################################
#' @title Plot Method for Local Polynomial Density Estimation and Inference
#'
#' @description The plot method for local polynomial density objects.
#'
#' @param ... Class "lpdensity" object, obtained from calling \code{\link{lpdensity}}.
#' @param alpha Numeric scalar between 0 and 1, specifies the significance level for plotting
#' confidence intervals/bands. If more than one is provided, they will be applied to each data series
#' accordingly.
#' @param type String, one of \code{"line"} (default), \code{"points"} and \code{"both"}, specifies how
#' the point estimates are plotted. If more than one is provided, they will be applied to each data series
#' accordingly.
#' @param lty Line type for point estimates, only effective if \code{type} is \code{"line"} or
#' \code{"both"}. \code{1} for solid line, \code{2} for dashed line, \code{3} for dotted line.
#' For other options, see the instructions for \code{ggplot2} or \code{\link{par}}. If
#' more than one is provided, they will be applied to each data series accordingly.
#' @param lwd Line width for point estimates, only effective if \code{type} is \code{"line"} or
#' \code{"both"}. Should be strictly positive. For other options, see the instructions for
#' \code{ggplot2} or \code{\link{par}}. If more than one is provided, they will be applied
#' to each data series accordingly.
#' @param lcol Line color for point estimates, only effective if \code{type} is \code{"line"} or
#' \code{"both"}. \code{1} for black, \code{2} for red, \code{3} for green, \code{4} for blue.
#' For other options, see the instructions for \code{ggplot2} or \code{\link{par}}. If
#' more than one is provided, they will be applied to each data series
#' accordingly.
#' @param pty Scatter plot type for point estimates, only effective if \code{type} is \code{"points"} or
#' \code{"both"}. For options, see the instructions for \code{ggplot2} or \code{\link{par}}. If
#' more than one is provided, they will be applied to each data series
#' accordingly.
#' @param pwd Scatter plot size for point estimates, only effective if \code{type} is \code{"points"} or
#' \code{"both"}. Should be strictly positive. If more than one is provided, they will be applied to each data series
#' accordingly.
#' @param pcol Scatter plot color for point estimates, only effective if \code{type} is \code{"points"} or
#' \code{"both"}. \code{1} for black, \code{2} for red, \code{3}
#' for green, \code{4} for blue.
#' For other options, see the instructions for \code{ggplot2} or \code{\link{par}}. If
#' more than one is provided, they will be applied to each data series
#' accordingly.
#' @param grid Numeric vector, specifies a subset of grid points
#' to plot point estimates. This option is effective only if \code{type} is \code{"points"} or
#' \code{"both"}; or if \code{CItype} is \code{"ebar"} or
#' \code{"all"}.
#' @param CItype String, one of \code{"region"} (shaded region, default), \code{"line"} (dashed lines),
#' \code{"ebar"} (error bars), \code{"all"} (all of the previous) or \code{"none"} (no confidence region),
#' how the confidence region should be plotted. If more than one is provided, they will be applied to each data series
#' accordingly.
#' @param CIuniform \code{TRUE} or \code{FALSE} (default), plotting either pointwise confidence intervals (\code{FALSE}) or
#' uniform confidence bands (\code{TRUE}).
#' @param CIsimul Positive integer, specifies the number of simulations used to construct critical values (default is \code{2000}). This
#' option is ignored if \code{CIuniform=FALSE}.
#' @param CIshade Numeric, specifies the opaqueness of the confidence region, should be between 0 (transparent) and
#' 1. Default is 0.2. If more than one is provided, they will be applied to each data series
#' accordingly.
#' @param CIcol Color of the confidence region. \code{1} for black, \code{2} for red, \code{3}
#' for green, \code{4} for blue.
#' For other options, see the instructions for \code{ggplot2} or \code{\link{par}}. If
#' more than one is provided, they will be applied to each data series
#' accordingly.
#' @param hist \code{TRUE} or \code{FALSE} (default), specifies whether a histogram should be added to the background.
#' @param histData Numeric vector, specifies the data used to construct the histogram plot.
#' @param histBreaks Numeric vector, specifies the breakpoints between histogram cells.
#' @param histFillCol Color of the histogram cells.
#' @param histFillShade Opaqueness of the histogram cells, should be between 0 (transparent) and
#' 1. Default is 0.2.
#' @param histLineCol Color of the histogram lines.
#' @param title,xlabel,ylabel Strings, specifies the title of the plot and labels for the x- and y-axis.
#' @param legendTitle String, specifies the legend title.
#' @param legendGroups String vector, specifies the group names used in legend.
#'
#' @return
#' A stadnard \code{ggplot} object is returned, hence can be used for further customization.
#'
#' @author
#' Matias D. Cattaneo, Princeton University. \email{cattaneo@princeton.edu}.
#'
#' Michael Jansson, University of California Berkeley. \email{mjansson@econ.berkeley.edu}.
#'
#' Xinwei Ma (maintainer), University of California San Diego. \email{x1ma@ucsd.edu}.
#'
#' @seealso \code{\link{lpdensity}} for local polynomial density estimation.
#'
#' Supported methods: \code{\link{coef.lpdensity}}, \code{\link{confint.lpdensity}},
#' \code{\link{plot.lpdensity}}, \code{\link{print.lpdensity}}, \code{\link{summary.lpdensity}},
#' \code{\link{vcov.lpdensity}}.
#'
#' @examples
#' # Generate a random sample
#' set.seed(42); X <- rnorm(2000)
#'
#' # Generate a density discontinuity at 0
#' X <- X - 0.5; X[X>0] <- X[X>0] * 2
#'
#' # Density estimation, left of 0 (scaled by the relative sample size)
#' est1 <- lpdensity(data = X[X<=0], grid = seq(-2.5, 0, 0.05), bwselect = "imse-dpi",
#' scale = sum(X<=0)/length(X))
#' # Density estimation, right of 0 (scaled by the relative sample size)
#' est2 <- lpdensity(data = X[X>0], grid = seq(0, 2, 0.05), bwselect = "imse-dpi",
#' scale = sum(X>0)/length(X))
#'
#' # Plot
#' plot(est1, est2, legendTitle="My Plot", legendGroups=c("Left", "Right"))
#'
#' # Plot uniform confidence bands
#' set.seed(42) # fix the seed for simulating critical values
#' plot(est1, est2, legendTitle="My Plot", legendGroups=c("Left", "Right"), CIuniform=TRUE)
#'
#' # Adding a histogram to the background
#' plot(est1, est2, legendTitle="My Plot", legendGroups=c("Left", "Right"),
#' hist=TRUE, histBreaks=seq(-2.4, 2, 0.2), histData=X)
#'
#' # Plot point estimates for a subset of evaluation points
#' plot(est1, est2, legendTitle="My Plot", legendGroups=c("Left", "Right"),
#' type="both", CItype="all", grid=seq(-2, 2, 0.5))
#' @export
plot.lpdensity <- function(..., alpha=NULL,
type=NULL, lty=NULL, lwd=NULL, lcol=NULL, pty=NULL, pwd=NULL, pcol=NULL, grid=NULL,
CItype=NULL, CIuniform=FALSE, CIsimul=2000, CIshade=NULL, CIcol=NULL,
hist=FALSE, histData=NULL, histBreaks=NULL, histFillCol=3, histFillShade=0.2, histLineCol="white",
title=NULL, xlabel=NULL, ylabel=NULL, legendTitle=NULL, legendGroups=NULL) {
########################################
# check how many series are passed in
########################################
x <- list(...)
nfig <- length(x)
if (nfig == 0) stop("Nothing to plot.\n")
flagToPlot <- rep(TRUE, nfig)
for (i in 1:length(x)) {
# check if is a lpbwdensity object
if (class(x[[i]])[1] == "lpbwdensity") {
flagToPlot[i] <- FALSE
warning(paste("Input ", i, " is an \"lpbwdensity\" object, which is not supported by the plot method.\n", sep=""))
next
}
# check if there is only one grid point
if (nrow(x[[i]]$Estimate) < 2) {
flagToPlot[i] <- FALSE
warning(paste("At least two grid points are needed to plot input ", i, ".\n", sep=""))
next
}
}
x <- x[flagToPlot]
nfig <- length(x)
if (nfig == 0) stop("Nothing to plot.\n")
isLpbwdensity <- FALSE
for (i in 1:nfig) {
if (class(x[[i]])[1] == "lpbwdensity") isLpbwdensity <- TRUE
}
if (isLpbwdensity) stop("The plot method does not support \"lpbwdensity\" objects.\n")
########################################
# error handling
########################################
# alpha
if (length(alpha) == 0) {
alpha <- rep(0.05, nfig)
} else if (!all(alpha>0 & alpha<1)) {
stop("Significance level incorrectly specified.\n")
} else {
alpha <- rep(alpha, length.out=nfig)
}
# plot type
if (length(type) == 0) {
type <- rep("line", nfig)
} else {
if (!all(type%in%c("line", "points", "both"))) {
stop("Plotting type incorrectly specified.\n")
}
type <- rep(type, length.out=nfig)
}
# CI type
if (length(CItype) == 0) {
CItype <- rep("region", nfig)
} else {
if (!all(CItype%in%c("region", "line", "ebar", "all", "none"))) {
stop("Confidence interval type incorrectly specified.\n")
}
CItype <- rep(CItype, length.out=nfig)
}
# line style, line width, line color
if (length(lty) == 0) {
lty <- rep(1, nfig)
} else {
lty <- rep(lty, length.out=nfig)
}
if (length(lwd) == 0) {
lwd <- rep(0.5, nfig)
} else {
lwd <- rep(lwd, length.out=nfig)
}
if (length(lcol) == 0) {
lcol <- 1:nfig
} else {
lcol <- rep(lcol, length.out=nfig)
}
# point style, point width, point color
if (length(pty) == 0) {
pty <- rep(1, nfig)
} else {
pty <- rep(pty, length.out=nfig)
}
if (length(pwd) == 0) {
pwd <- rep(1, nfig)
} else {
pwd <- rep(pwd, length.out=nfig)
}
if (length(pcol) == 0) {
pcol <- lcol
} else {
pcol <- rep(pcol, length.out=nfig)
}
# CI shade, CI color
if (length(CIshade) == 0) {
CIshade <- rep(0.2, nfig)
} else {
CIshade <- rep(CIshade, length.out=nfig)
}
if (length(CIcol) == 0) {
CIcol <- lcol
} else {
CIcol <- rep(CIcol, length.out=nfig)
}
# legend
# New in v0.2.1 to handle legend
if (length(legendTitle) == 0) {
legendTitle <- ""
} else {
legendTitle <- legendTitle[1]
}
if (length(legendGroups) > 0) {
legendGroups <- rep(legendGroups, length.out=nfig)
legend_default <- FALSE
} else {
legend_default <- TRUE
}
# grid
if (!is.null(grid)) {
if (!is.numeric(grid)) {
stop("Option grid incorrectly specified.\n")
}
}
########################################
# initializing plot
########################################
if (hist & !is.null(histData)) {
histData <- as.data.frame(histData)
colnames(histData) <- c("v1")
if(is.null(histBreaks)) { histBreaks <- seq(from=min(histData[, 1]), to=max(histData[, 1]), length.out=21) }
histScale <- mean(histData[, 1] >= min(histBreaks) & histData[, 1] <= max(histBreaks))
temp_plot <- ggplot() +
geom_histogram(data=histData, aes(x=v1, y=after_stat(density)*histScale), breaks=histBreaks, fill=histFillCol, col=histLineCol, alpha=histFillShade) +
theme_bw() #+ theme(legend.position="none")
} else {
temp_plot <- ggplot() + theme_bw() #+ theme(legend.position="none")
}
#CI_l <- CI_r <- f_p <- Sname <- v1 <- ..density.. <- NULL
CI_l <- CI_r <- f_p <- Sname <- v1 <- NULL
########################################
# looping over input models
########################################
### all colors
col_all <- lty_all <- pty_all <- v_all <- c()
estRangeL <- estRangeR <- c() # estimation range
for (i in 1:nfig) {
# get derivative order
v_all <- c(v_all, x[[i]]$opt$v)
# get ploting indices
if (is.null(grid)) {
plotIndex <- 1:nrow(x[[i]]$Estimate)
} else {
gridTemp <- grid[grid >= min(x[[i]]$Estimate[, "grid"]) & grid <= max(x[[i]]$Estimate[, "grid"])]
if (length(gridTemp) == 0) {
plotIndex <- NULL
} else {
plotIndex <- rep(NA, length(gridTemp))
for (ii in 1:length(gridTemp)) {
plotIndex[ii] <- which.min(abs(gridTemp[ii]-x[[i]]$Estimate[, "grid"]))
}
plotIndex <- unique(plotIndex)
}
}
estRangeL <- min(estRangeL, min(x[[i]]$Estimate[, "grid"]))
estRangeR <- max(estRangeR, max(x[[i]]$Estimate[, "grid"]))
data_x <- data.frame(x[[i]]$Estimate[, c("grid", "f_p", "f_q", "se_p", "se_q"), drop=FALSE])
if (x[[i]]$opt$q == x[[i]]$opt$p) {
data_x$f_q <- data_x$f_p; data_x$se_q <- data_x$se_p
}
# critical value
if (CIuniform) {
if (length(CIsimul) == 0) { CIsimul <- 2000 }
if (!is.numeric(CIsimul) | is.na(CIsimul)) {
warning("Option CIsimul incorrectly specified. Will only plot pointwise confidence intervals.\n")
z_val <- qnorm(1 - alpha[i]/2)
} else if (ceiling(CIsimul)<2) {
warning("Option CIsimul incorrectly specified. Will only plot pointwise confidence intervals.\n")
z_val <- qnorm(1 - alpha[i]/2)
} else {
CIsimul <- ceiling(CIsimul)
if (x[[i]]$opt$q == x[[i]]$opt$p) {
x[[i]]$CovMat_q <- x[[i]]$CovMat_p
}
corrMat <- sweep(sweep(x[[i]]$CovMat_q, MARGIN=1, FUN="*", STATS=1/data_x$se_q), MARGIN=2, FUN="*", STATS=1/data_x$se_q)
normalSimu <- try(
mvrnorm(n=CIsimul, mu=rep(0,nrow(corrMat)), Sigma=corrMat),
silent=TRUE)
if (is.character(normalSimu)) {
print(normalSimu)
warning("Variance-Covariance is not positive semidefinite. Will only plot pointwise confidence intervals.\n")
z_val <- qnorm(1 - alpha[i]/2)
} else {
z_val <- quantile(apply(normalSimu, MARGIN=1, FUN=function(x) {max(abs(x))}), 1 - alpha[i])
}
}
} else {
z_val <- qnorm(1 - alpha[i]/2)
}
data_x$CI_l <- data_x$f_q - z_val * data_x$se_q
data_x$CI_r <- data_x$f_q + z_val * data_x$se_q
# New in v0.2.1 to handle legend
if (legend_default) {
data_x$Sname <- paste("Series", i, sep=" ")
legendGroups <- c(legendGroups, data_x$Sname)
} else {
data_x$Sname <- legendGroups[i]
}
########################################
# add CI regions to the plot
if (CItype[i]%in%c("region", "all"))
temp_plot <- temp_plot + geom_ribbon(data=data_x, aes(x=grid, ymin=CI_l, ymax=CI_r), alpha=CIshade[i], fill=CIcol[i])
########################################
# add CI lines to the plot
if (CItype[i]%in%c("line", "all"))
temp_plot <- temp_plot + geom_line(data=data_x, aes(x=grid, y=CI_l), linetype=2, alpha=1, col=CIcol[i]) +
geom_line(data=data_x, aes(x=grid, y=CI_r), linetype=2, alpha=1, col=CIcol[i])
#if (CItype[i]%in%c("line", "all"))
# temp_plot <- temp_plot + geom_line(data=data_x, aes(x=grid, y=CI_l), linetype=2, alpha=CIshade[i], col=CIcol[i]) +
# geom_line(data=data_x, aes(x=grid, y=CI_r), linetype=2, alpha=CIshade[i], col=CIcol[i])
########################################
# add error bars to the plot
if (CItype[i]%in%c("ebar", "all") & !is.null(plotIndex))
temp_plot <- temp_plot + geom_errorbar(data=data_x[plotIndex, ], aes(x=grid, ymin=CI_l, ymax=CI_r), alpha=1, col=CIcol[i], linetype=1)
#if (CItype[i]%in%c("ebar", "all") & !is.null(plotIndex))
# temp_plot <- temp_plot + geom_errorbar(data=data_x[plotIndex, ], aes(x=grid, ymin=CI_l, ymax=CI_r), alpha=CIshade[i], col=CIcol[i], linetype=1)
########################################
# add lines to the plot
if (type[i]%in%c("line", "both")) {
temp_plot <- temp_plot + geom_line(data=data_x, aes(x=grid, y=f_p, colour=Sname, linetype=Sname), size=lwd[i])
}
########################################
# add points to the plot
if (type[i]%in%c("points", "both") & !is.null(plotIndex)) {
temp_plot <- temp_plot + geom_point(data=data_x[plotIndex, ], aes(x=grid, y=f_p, colour=Sname, shape=Sname), size=pwd[i])
}
if (type[i] == "line") {
col_all <- c(col_all, lcol[i])
lty_all <- c(lty_all, lty[i])
pty_all <- c(pty_all, NA)
} else if (type[i] == "both") {
col_all <- c(col_all, lcol[i])
lty_all <- c(lty_all, lty[i])
pty_all <- c(pty_all, pty[i])
} else {
col_all <- c(col_all, pcol[i])
lty_all <- c(lty_all, NA)
pty_all <- c(pty_all, pty[i])
}
}
########################################
# change color, line type and point shape back, and customize legend
########################################
# New in v0.2.1 to handle legend
index <- sort.int(legendGroups, index.return=TRUE)$ix
temp_plot <- temp_plot + scale_color_manual(values = col_all[index]) +
scale_linetype_manual(values = lty_all[index]) +
scale_shape_manual(values = pty_all[index]) +
guides(colour=guide_legend(title=legendTitle)) +
guides(linetype=guide_legend(title=legendTitle)) +
guides(shape=guide_legend(title=legendTitle))
########################################
# add title, x and y labs
########################################
if (is.null(ylabel)) {
if (all(v_all == v_all[1])) {
if (v_all[1] == 0) {
ylabel <- "Distribution function"
} else if (v_all[1] == 1) {
ylabel <- "Density"
} else {
ylabel <- paste("Density derivative (v=", v_all[1], ")", sep="")
}
} else {
ylabel <- ""
}
}
if (is.null(xlabel)) {
xlabel <- ""
}
if (is.null(title)) {
title <- ""
}
temp_plot <- temp_plot + labs(x=xlabel, y=ylabel) + ggtitle(title)
# check plotting range vs estimation range
if (!is.null(grid)) {
if (min(grid) < estRangeL | max(grid) > estRangeR) {
warning("The plotting range exceeds the original estimation range. Option plot(..., grid=) should be within the estimation range specified by lpdensity(..., grid=).\n")
}
}
########################################
# return the plot
########################################
return (temp_plot)
}
################################################################################
#' @title Plot Method for Local Polynomial Density Estimation and Inference
#'
#' @description This has been replaced by \code{\link{plot.lpdensity}}.
#'
#' @param ... Class "lpdensity" object, obtained from calling \code{\link{lpdensity}}.
#' @param alpha Numeric scalar between 0 and 1, specifies the significance level for plotting
#' confidence intervals/bands. If more than one is provided, they will be applied to each data series
#' accordingly.
#' @param type String, one of \code{"line"} (default), \code{"points"} and \code{"both"}, specifies how
#' the point estimates are plotted. If more than one is provided, they will be applied to each data series
#' accordingly.
#' @param lty Line type for point estimates, only effective if \code{type} is \code{"line"} or
#' \code{"both"}. \code{1} for solid line, \code{2} for dashed line, \code{3} for dotted line.
#' For other options, see the instructions for \code{ggplot2} or \code{\link{par}}. If
#' more than one is provided, they will be applied to each data series accordingly.
#' @param lwd Line width for point estimates, only effective if \code{type} is \code{"line"} or
#' \code{"both"}. Should be strictly positive. For other options, see the instructions for
#' \code{ggplot2} or \code{\link{par}}. If more than one is provided, they will be applied
#' to each data series accordingly.
#' @param lcol Line color for point estimates, only effective if \code{type} is \code{"line"} or
#' \code{"both"}. \code{1} for black, \code{2} for red, \code{3} for green, \code{4} for blue.
#' For other options, see the instructions for \code{ggplot2} or \code{\link{par}}. If
#' more than one is provided, they will be applied to each data series
#' accordingly.
#' @param pty Scatter plot type for point estimates, only effective if \code{type} is \code{"points"} or
#' \code{"both"}. For options, see the instructions for \code{ggplot2} or \code{\link{par}}. If
#' more than one is provided, they will be applied to each data series
#' accordingly.
#' @param pwd Scatter plot size for point estimates, only effective if \code{type} is \code{"points"} or
#' \code{"both"}. Should be strictly positive. If more than one is provided, they will be applied to each data series
#' accordingly.
#' @param pcol Scatter plot color for point estimates, only effective if \code{type} is \code{"points"} or
#' \code{"both"}. \code{1} for black, \code{2} for red, \code{3}
#' for green, \code{4} for blue.
#' For other options, see the instructions for \code{ggplot2} or \code{\link{par}}. If
#' more than one is provided, they will be applied to each data series
#' accordingly.
#' @param grid Numeric vector, specifies a subset of grid points
#' to plot point estimates. This option is effective only if \code{type} is \code{"points"} or
#' \code{"both"}; or if \code{CItype} is \code{"ebar"} or
#' \code{"all"}.
#' @param CItype String, one of \code{"region"} (shaded region, default), \code{"line"} (dashed lines),
#' \code{"ebar"} (error bars), \code{"all"} (all of the previous) or \code{"none"} (no confidence region),
#' how the confidence region should be plotted. If more than one is provided, they will be applied to each data series
#' accordingly.
#' @param CIuniform \code{TRUE} or \code{FALSE} (default), plotting either pointwise confidence intervals (\code{FALSE}) or
#' uniform confidence bands (\code{TRUE}).
#' @param CIsimul Positive integer, specifies the number of simulations used to construct critical values (default is \code{2000}). This
#' option is ignored if \code{CIuniform=FALSE}.
#' @param CIshade Numeric, specifies the opaqueness of the confidence region, should be between 0 (transparent) and
#' 1. Default is 0.2. If more than one is provided, they will be applied to each data series
#' accordingly.
#' @param CIcol Color of the confidence region. \code{1} for black, \code{2} for red, \code{3}
#' for green, \code{4} for blue.
#' For other options, see the instructions for \code{ggplot2} or \code{\link{par}}. If
#' more than one is provided, they will be applied to each data series
#' accordingly.
#' @param hist \code{TRUE} or \code{FALSE} (default), specifies whether a histogram should be added to the background.
#' @param histData Numeric vector, specifies the data used to construct the histogram plot.
#' @param histBreaks Numeric vector, specifies the breakpoints between histogram cells.
#' @param histFillCol Color of the histogram cells.
#' @param histFillShade Opaqueness of the histogram cells, should be between 0 (transparent) and
#' 1. Default is 0.2.
#' @param histLineCol Color of the histogram lines.
#' @param title,xlabel,ylabel Strings, specifies the title of the plot and labels for the x- and y-axis.
#' @param legendTitle String, specifies the legend title.
#' @param legendGroups String vector, specifies the group names used in legend.
#'
#' @return
#' A stadnard \code{ggplot} object is returned, hence can be used for further customization.
#'
#' @author
#' Matias D. Cattaneo, Princeton University. \email{cattaneo@princeton.edu}.
#'
#' Michael Jansson, University of California Berkeley. \email{mjansson@econ.berkeley.edu}.
#'
#' Xinwei Ma (maintainer), University of California San Diego. \email{x1ma@ucsd.edu}.
#'
#' @export
lpdensity.plot <- plot.lpdensity
################################################################################
#' Coef Method for Local Polynomial Density Estimation and Inference
#'
#' @description The coef method for local polynomial density objects.
#'
#' @param object Class "lpdensity" object, obtained by calling \code{\link{lpdensity}}.
#' @param ... Additional options.
#'
#' @return
#' A matrix containing grid points and density estimates using p- and q-th order local polynomials.
#'
#' @author
#' Matias D. Cattaneo, Princeton University. \email{cattaneo@princeton.edu}.
#'
#' Michael Jansson, University of California Berkeley. \email{mjansson@econ.berkeley.edu}.
#'
#' Xinwei Ma (maintainer), University of California San Diego. \email{x1ma@ucsd.edu}.
#'
#' @seealso \code{\link{lpdensity}} for local polynomial density estimation.
#'
#' Supported methods: \code{\link{coef.lpdensity}}, \code{\link{confint.lpdensity}},
#' \code{\link{plot.lpdensity}}, \code{\link{print.lpdensity}}, \code{\link{summary.lpdensity}},
#' \code{\link{vcov.lpdensity}}.
#'
#' @examples
#' # Generate a random sample
#' set.seed(42); X <- rnorm(2000)
#'
#' # Estimate density and report results
#' coef(lpdensity(data = X, bwselect = "imse-dpi"))
#'
#' @export
coef.lpdensity <- function(object, ...) {
object$Estimate[, c("grid", "f_p", "f_q")]
}
################################################################################
#' Vcov Method for Local Polynomial Density Estimation and Inference
#'
#' @description The vcov method for local polynomial density objects.
#'
#' @param object Class "lpdensity" object, obtained by calling \code{\link{lpdensity}}.
#' @param ... Additional options.
#'
#' @return
#' \item{stdErr}{A matrix containing grid points and standard errors using p- and q-th order local polynomials.}
#' \item{CovMat_p}{The variance-covariance matrix corresponding to \code{f_p}.}
#' \item{CovMat_q}{The variance-covariance matrix corresponding to \code{f_q}.}
#'
#' @author
#' Matias D. Cattaneo, Princeton University. \email{cattaneo@princeton.edu}.
#'
#' Michael Jansson, University of California Berkeley. \email{mjansson@econ.berkeley.edu}.
#'
#' Xinwei Ma (maintainer), University of California San Diego. \email{x1ma@ucsd.edu}.
#'
#' @seealso \code{\link{lpdensity}} for local polynomial density estimation.
#'
#' Supported methods: \code{\link{coef.lpdensity}}, \code{\link{confint.lpdensity}},
#' \code{\link{plot.lpdensity}}, \code{\link{print.lpdensity}}, \code{\link{summary.lpdensity}},
#' \code{\link{vcov.lpdensity}}.
#'
#' @examples
#' # Generate a random sample
#' set.seed(42); X <- rnorm(2000)
#'
#' # Estimate density and report results
#' vcov(lpdensity(data = X, bwselect = "imse-dpi"))
#'
#' @export
vcov.lpdensity <- function(object, ...) {
if (class(object)[1] == "lpbwdensity") stop("The vcov method does not support \"lpbwdensity\" objects.\n")
return(list(stdErr=object$Estimate[, c("grid", "se_p", "se_q")], CovMat_p=object$CovMat_p, CovMat_q=object$CovMat_q))
}
################################################################################
#' Confint Method for Local Polynomial Density Estimation and Inference
#'
#' @description The confint method for local polynomial density objects.
#'
#' @param object Class "lpdensity" object, obtained by calling \code{\link{lpdensity}}.
#' @param parm Integer, indicating which parameters are to be given confidence intervals.
#' @param level Numeric scalar between 0 and 1, the significance level for computing confidence intervals
#' @param ... Additional options, including (i) \code{grid} specifies a subset of grid points
#' to display the bandwidth; (ii) \code{gridIndex} specifies the indices of grid points
#' to display the bandwidth (this is the same as \code{parm}); (iii) \code{alpha} specifies the significance level
#' (this is \code{1-level}); (iv)
#' \code{CIuniform} specifies whether displaying pointwise confidence intervals (\code{FALSE}, default) or
#' the uniform confidence band (\code{TRUE}); (v) \code{CIsimul} specifies the number of simulations used
#' to construct critical values (default is 2000).
#'
#' @return
#' A matrix containing grid points and confidence interval end points using p- and q-th order local polynomials.
#'
#' @author
#' Matias D. Cattaneo, Princeton University. \email{cattaneo@princeton.edu}.
#'
#' Michael Jansson, University of California Berkeley. \email{mjansson@econ.berkeley.edu}.
#'
#' Xinwei Ma (maintainer), University of California San Diego. \email{x1ma@ucsd.edu}.
#'
#' @seealso \code{\link{lpdensity}} for local polynomial density estimation.
#'
#' Supported methods: \code{\link{coef.lpdensity}}, \code{\link{confint.lpdensity}},
#' \code{\link{plot.lpdensity}}, \code{\link{print.lpdensity}}, \code{\link{summary.lpdensity}},
#' \code{\link{vcov.lpdensity}}.
#'
#' @examples
#' # Generate a random sample
#' set.seed(42); X <- rnorm(2000)
#'
#' # Estimate density and report 95% confidence intervals
#' est1 <- lpdensity(data = X, bwselect = "imse-dpi")
#' confint(est1)
#'
#' # Report results for a subset of grid points
#' confint(est1, parm=est1$Estimate[4:10, "grid"])
#' confint(est1, grid=est1$Estimate[4:10, "grid"])
#' confint(est1, gridIndex=4:10)
#'
#' # Report the 99% uniform confidence band
#' # Fix the seed for simulating critical values
#' set.seed(42); confint(est1, level=0.99, CIuniform=TRUE)
#' set.seed(42); confint(est1, alpha=0.01, CIuniform=TRUE)
#'
#' @export
confint.lpdensity <- function(object, parm = NULL, level = NULL, ...) {
x <- object
if (class(x)[1] == "lpbwdensity") stop("The confint method does not support \"lpbwdensity\" objects.\n")
args <- list(...)
if (!is.null(parm)) { args[['grid']] <- parm }
if (!is.null(level)) { args[['alpha']] <- 1 - level }
if (is.null(args[['alpha']])) { alpha <- 0.05 } else { alpha <- args[['alpha']] }
if (is.null(args[['CIuniform']])) { CIuniform <- FALSE } else { CIuniform <- args[['CIuniform']] }
if (is.null(args[['CIsimul']])) { CIsimul <- 2000 } else { sep <- args[['CIsimul']] }
if (is.null(args[['grid']]) & is.null(args[['gridIndex']])) {
gridIndex <- 1:nrow(x$Estimate)
} else if (is.null(args[['grid']]) & !is.null(args[['gridIndex']])) {
gridIndex <- args[['gridIndex']]
if (is.null(gridIndex)) {
gridIndex <- 1:nrow(x$Estimate)
} else if (!all(gridIndex %in% 1:nrow(x$Estimate))) {
stop("Option gridIndex incorrectly specified.\n")
}
} else {
grid <- args[['grid']]
if (is.null(grid)) {
gridIndex <- 1:nrow(x$Estimate)
} else if (!is.numeric(grid)) {
stop("Option param/grid incorrectly specified.\n")
} else {
gridIndex <- rep(NA, length(grid))
for (j in 1:length(grid)) {
gridIndex[j] <- which.min(abs(x$Estimate[, "grid"]-grid[j]))
#gridIndex <- unique(gridIndex)
}
}
}
Estimate <- matrix(NA, nrow=length(gridIndex), ncol=5)
Estimate[, 1] <- x$Estimate[gridIndex, "grid"]
if (CIuniform) {
if (length(CIsimul) == 0) { CIsimul <- 2000 }
if (!is.numeric(CIsimul) | is.na(CIsimul)) {
warning("Option CIsimul incorrectly specified. Will only plot pointwise confidence intervals.\n")
CIuniform <- FALSE
z_val <- qnorm(1 - alpha/2)
} else if (ceiling(CIsimul)<2) {
warning("Option CIsimul incorrectly specified. Will only plot pointwise confidence intervals.\n")
CIuniform <- FALSE
z_val <- qnorm(1 - alpha/2)
} else {
CIsimul <- ceiling(CIsimul)
if (x$opt$q == x$opt$p) {
x$CovMat_q <- x$CovMat_p
x$Estimate[, "se_q"] <- x$Estimate[, "se_p"]
}
corrMat <- sweep(sweep(x$CovMat_q, MARGIN=1, FUN="*", STATS=1/x$Estimate[, "se_q"]), MARGIN=2, FUN="*", STATS=1/x$Estimate[, "se_q"])
normalSimu <- try(
mvrnorm(n=CIsimul, mu=rep(0,nrow(corrMat)), Sigma=corrMat),
silent=TRUE)
if (is.character(normalSimu)) {
print(normalSimu)
warning("Variance-Covariance is not positive semidefinite. Will only plot pointwise confidence intervals.\n")
CIuniform <- FALSE
z_val <- qnorm(1 - alpha/2)
} else {
z_val <- quantile(apply(normalSimu, MARGIN=1, FUN=function(x) {max(abs(x))}), 1 - alpha)
}
}
} else {
z_val <- qnorm(1 - alpha/2)
}
Estimate[, 2] <- x$Estimate[gridIndex, "f_p"] - z_val * x$Estimate[gridIndex, "se_p"]
Estimate[, 3] <- x$Estimate[gridIndex, "f_p"] + z_val * x$Estimate[gridIndex, "se_p"]
Estimate[, 4] <- x$Estimate[gridIndex, "f_q"] - z_val * x$Estimate[gridIndex, "se_q"]
Estimate[, 5] <- x$Estimate[gridIndex, "f_q"] + z_val * x$Estimate[gridIndex, "se_q"]
colnames(Estimate) <- c("grid", "CI_l_p", "CI_r_p", "CI_l_q", "CI_r_q")
#if (x$opt$p == x$opt$q) Estimate <- Estimate[, 1:3]
return(Estimate)
}
Try the lpdensity package in your browser
Any scripts or data that you put into this service are public.
lpdensity documentation built on Oct. 6, 2024, 9:06 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions confint.lpdensity vcov.lpdensity coef.lpdensity plot.lpdensity summary.lpdensity print.lpdensity
Documented in coef.lpdensity confint.lpdensity plot.lpdensity print.lpdensity summary.lpdensity vcov.lpdensity
################################################################################
#' Print Method for Local Polynomial Density Estimation and Inference
#'
#' @description The print method for local polynomial density objects.
#'
#' @param x Class "lpdensity" object, obtained from calling \code{\link{lpdensity}}.
#' @param ... Additional options.
#'
#' @author
#' Matias D. Cattaneo, Princeton University. \email{cattaneo@princeton.edu}.
#'
#' Michael Jansson, University of California Berkeley. \email{mjansson@econ.berkeley.edu}.
#'
#' Xinwei Ma (maintainer), University of California San Diego. \email{x1ma@ucsd.edu}.
#'
#' @seealso \code{\link{lpdensity}} for local polynomial density estimation.
#'
#' Supported methods: \code{\link{coef.lpdensity}}, \code{\link{confint.lpdensity}},
#' \code{\link{plot.lpdensity}}, \code{\link{print.lpdensity}}, \code{\link{summary.lpdensity}},
#' \code{\link{vcov.lpdensity}}.
#'
#' @examples
#' # Generate a random sample
#' set.seed(42); X <- rnorm(2000)
#'
#' # Estimate density and report results
#' print(lpdensity(data = X, bwselect = "imse-dpi"))
#'
#' @export
print.lpdensity <- function(x, ...) {
cat("Call: lpdensity\n\n")
cat(paste("Sample size ", x$opt$n, "\n", sep=""))
cat(paste("Polynomial order for point estimation (p=) ", x$opt$p, "\n", sep=""))
cat(paste("Order of derivative estimated (v=) ", x$opt$v, "\n", sep=""))
if (x$opt$q > 0) {
cat(paste("Polynomial order for confidence interval (q=) ", x$opt$q, "\n", sep=""))
} else {
cat(paste("Polynomial order for confidence interval (q=) ", "p", "\n", sep=""))
}
cat(paste("Kernel function ", x$opt$kernel, "\n", sep=""))
if (x$opt$scale != 1) {
cat(paste("Scaling factor ", x$opt$scale, "\n", sep=""))
}
cat(paste("Bandwidth method ", x$opt$bwselect, "\n", sep=""))
cat("\n")
cat("Use summary(...) to show estimates.\n")
}
################################################################################
#' Summary Method for Local Polynomial Density Estimation and Inference
#'
#' @description The summary method for local polynomial density objects.
#'
#' @param object Class "lpdensity" object, obtained from calling \code{\link{lpdensity}}.
#' @param ... Additional options, including (i) \code{grid} specifies a subset of grid points
#' to display results; (ii) \code{gridIndex} specifies the indices of grid points
#' to display results; (iii) \code{alpha} specifies the significance level; (iv)
#' \code{CIuniform} specifies whether displaying pointwise confidence intervals (\code{FALSE}, default) or
#' the uniform confidence band (\code{TRUE}); (v) \code{CIsimul} specifies the number of simulations used
#' to construct critical values (default is \code{2000}).
#'
#' @author
#' Matias D. Cattaneo, Princeton University. \email{cattaneo@princeton.edu}.
#'
#' Michael Jansson, University of California Berkeley. \email{mjansson@econ.berkeley.edu}.
#'
#' Xinwei Ma (maintainer), University of California San Diego. \email{x1ma@ucsd.edu}.
#'
#' @seealso \code{\link{lpdensity}} for local polynomial density estimation.
#'
#' Supported methods: \code{\link{coef.lpdensity}}, \code{\link{confint.lpdensity}},
#' \code{\link{plot.lpdensity}}, \code{\link{print.lpdensity}}, \code{\link{summary.lpdensity}},
#' \code{\link{vcov.lpdensity}}.
#'
#' @examples
#' # Generate a random sample
#' set.seed(42); X <- rnorm(2000)
#'
#' # Estimate density and report results
#' est1 <- lpdensity(data = X, bwselect = "imse-dpi")
#' summary(est1)
#'
#' # Report results for a subset of grid points
#' summary(est1, grid=est1$Estimate[4:10, "grid"])
#' summary(est1, gridIndex=4:10)
#'
#' # Report the 99% uniform confidence band
#' set.seed(42) # fix the seed for simulating critical values
#' summary(est1, alpha=0.01, CIuniform=TRUE)
#'
#' @export
summary.lpdensity <- function(object, ...) {
x <- object
args <- list(...)
if (is.null(args[['alpha']])) { alpha <- 0.05 } else { alpha <- args[['alpha']] }
if (is.null(args[['sep']])) { sep <- 5 } else { sep <- args[['sep']] }
if (is.null(args[['CIuniform']])) { CIuniform <- FALSE } else { CIuniform <- args[['CIuniform']] }
if (is.null(args[['CIsimul']])) { CIsimul <- 2000 } else { sep <- args[['CIsimul']] }
if (is.null(args[['grid']]) & is.null(args[['gridIndex']])) {
gridIndex <- 1:nrow(x$Estimate)
} else if (is.null(args[['grid']]) & !is.null(args[['gridIndex']])) {
gridIndex <- args[['gridIndex']]
if (is.null(gridIndex)) {
gridIndex <- 1:nrow(x$Estimate)
} else if (!all(gridIndex %in% 1:nrow(x$Estimate))) {
stop(paste("Option gridIndex incorrectly specified. Should be integers between 1 and ", nrow(x$Estimate), ".\n", sep=""))
}
} else {
grid <- args[['grid']]
if (is.null(grid)) {
gridIndex <- 1:nrow(x$Estimate)
} else if (!is.numeric(grid)) {
stop("Option grid incorrectly specified.\n")
} else {
gridIndex <- rep(NA, length(grid))
if (min(grid) < min(x$Estimate[, "grid"]) | max(grid) > max(x$Estimate[, "grid"])) {
warning("The reporting range exceeds the original estimation range. Option summary(..., grid=) should be within the estimation range specified by lpdensity(..., grid=).\n")
}
for (j in 1:length(grid)) {
gridIndex[j] <- which.min(abs(x$Estimate[, "grid"]-grid[j]))
#gridIndex <- unique(gridIndex)
}
}
}
cat("Call: lpdensity\n\n")
cat(paste("Sample size (n=) ", x$opt$n, "\n", sep=""))
cat(paste("Polynomial order for point estimation (p=) ", x$opt$p, "\n", sep=""))
if (x$opt$v == 0) {
cat(paste("Distribution function estimated (v=) ", x$opt$v, "\n", sep=""))
} else if (x$opt$v == 1) {
cat(paste("Density function estimated (v=) ", x$opt$v, "\n", sep=""))
} else {
cat(paste("Order of derivative estimated (v=) ", x$opt$v, "\n", sep=""))
}
if (x$opt$q > x$opt$p) {
cat(paste("Polynomial order for confidence interval (q=) ", x$opt$q, "\n", sep=""))
} else {
cat(paste("Polynomial order for confidence interval (q=) ", "p", "\n", sep=""))
}
cat(paste("Kernel function ", x$opt$kernel, "\n", sep=""))
if (x$opt$scale != 1) {
cat(paste("Scaling factor ", x$opt$scale, "\n", sep=""))
}
cat(paste("Bandwidth selection method ", x$opt$bwselect, "\n", sep=""))
cat("\n")
### compute CI
if (CIuniform) {
if (length(CIsimul) == 0) { CIsimul <- 2000 }
if (!is.numeric(CIsimul) | is.na(CIsimul)) {
warning("Option CIsimul incorrectly specified. Will only plot pointwise confidence intervals.\n")
CIuniform <- FALSE
z_val <- qnorm(1 - alpha/2)
} else if (ceiling(CIsimul)<2) {
warning("Option CIsimul incorrectly specified. Will only plot pointwise confidence intervals.\n")
CIuniform <- FALSE
z_val <- qnorm(1 - alpha/2)
} else {
CIsimul <- ceiling(CIsimul)
if (x$opt$q == x$opt$p) {
x$CovMat_q <- x$CovMat_p
x$Estimate[, "se_q"] <- x$Estimate[, "se_p"]
}
corrMat <- sweep(sweep(x$CovMat_q, MARGIN=1, FUN="*", STATS=1/x$Estimate[, "se_q"]), MARGIN=2, FUN="*", STATS=1/x$Estimate[, "se_q"])
normalSimu <- try(
mvrnorm(n=CIsimul, mu=rep(0,nrow(corrMat)), Sigma=corrMat),
silent=TRUE)
if (is.character(normalSimu)) {
print(normalSimu)
warning("Variance-Covariance is not positive semidefinite. Will only plot pointwise confidence intervals.\n")
CIuniform <- FALSE
z_val <- qnorm(1 - alpha/2)
} else {
z_val <- quantile(apply(normalSimu, MARGIN=1, FUN=function(x) {max(abs(x))}), 1 - alpha)
}
}
} else {
z_val <- qnorm(1 - alpha/2)
}
if (x$opt$q == x$opt$p) {
CI_l <- x$Estimate[, "f_p"] - x$Estimate[, "se_p"] * z_val
CI_r <- x$Estimate[, "f_p"] + x$Estimate[, "se_p"] * z_val
} else {
CI_l <- x$Estimate[, "f_q"] - x$Estimate[, "se_q"] * z_val
CI_r <- x$Estimate[, "f_q"] + x$Estimate[, "se_q"] * z_val
}
flagNotAllLocalSampleSizeSame <- !all(x$Estimate[, "nh"] == x$Estimate[, "nhu"])
### print output
cat(paste(rep("=", 14 + 10 + 8 + 8*flagNotAllLocalSampleSizeSame + 10 + 10 + 25), collapse="")); cat("\n")
cat(format(" ", width= 14 ))
cat(format(" ", width= 10 ))
cat(format(" ", width= 8 + 8*flagNotAllLocalSampleSizeSame ))
cat(format("Point", width= 10, justify="right"))
cat(format("Std." , width= 10, justify="right"))
if (x$opt$q > x$opt$p) {
cat(format("Robust B.C."
, width=25, justify="centre"))
cat("\n")
} else {
cat(format("Conventional"
, width=25, justify="centre"))
cat("\n")
}
cat(format("Index Grid" , width=14, justify="right"))
cat(format("B.W." , width=10, justify="right"))
cat(format("Eff.n" , width=8 , justify="right"))
if (flagNotAllLocalSampleSizeSame) {
cat(format("Uniq.n" , width=8 , justify="right"))
}
cat(format("Est." , width=10, justify="right"))
cat(format("Error" , width=10, justify="right"))
if (CIuniform) {
cat(format(paste("[ Unif. ", floor((1-alpha)*100), "%", " C.I. ]", sep="")
, width=25, justify="centre"))
} else {
cat(format(paste("[ ", floor((1-alpha)*100), "%", " C.I. ]", sep="")
, width=25, justify="centre"))
}
cat("\n")
if (all(x$Estimate[, "nh"] == x$Estimate[, "nhu"])) {
cat(paste(rep("=", 14 + 10 + 8 + 10 + 10 + 25), collapse="")); cat("\n")
jj <- 1
for (j in gridIndex) {
cat(format(toString(j), width=4))
cat(format(sprintf("%6.4f", x$Estimate[j, "grid"]), width=10, justify="right"))
cat(format(sprintf("%6.4f", x$Estimate[j, "bw"]) , width=10, justify="right"))
cat(format(sprintf("%8.0f", x$Estimate[j, "nh"]) , width=8 , justify="right"))
cat(format(sprintf("%6.4f", x$Estimate[j, "f_p"]) , width=10, justify="right"))
cat(format(
paste(sprintf("%6.4f", x$Estimate[j, "se_p"]), sep=""), width=10, justify="right"))
cat(format(
paste(sprintf("%7.4f", CI_l[j]), " , ", sep="") , width=14, justify="right"))
cat(format(
paste(sprintf("%7.4f", CI_r[j]), sep=""), width=11, justify="left"))
cat("\n")
if (is.numeric(sep)) if (sep > 0) if (jj %% sep == 0) {
cat(paste(rep("-", 14 + 10 + 8 + 10 + 10 + 25), collapse="")); cat("\n")
}
jj <- jj + 1
}
cat(paste(rep("=", 14 + 10 + 8 + 10 + 10 + 25), collapse="")); cat("\n")
} else {
cat(paste(rep("=", 14 + 10 + 8 + 8 + 10 + 10 + 25), collapse="")); cat("\n")
jj <- 1
for (j in gridIndex) {
cat(format(toString(j), width=4))
cat(format(sprintf("%6.4f", x$Estimate[j, "grid"]), width=10, justify="right"))
cat(format(sprintf("%6.4f", x$Estimate[j, "bw"]) , width=10, justify="right"))
cat(format(sprintf("%8.0f", x$Estimate[j, "nh"]) , width=8 , justify="right"))
cat(format(sprintf("%8.0f", x$Estimate[j, "nhu"]) , width=8 , justify="right"))
cat(format(sprintf("%6.4f", x$Estimate[j, "f_p"]) , width=10, justify="right"))
cat(format(
paste(sprintf("%6.4f", x$Estimate[j, "se_p"]), sep=""), width=10, justify="right"))
cat(format(
paste(sprintf("%7.4f", CI_l[j]), " , ", sep="") , width=14, justify="right"))
cat(format(
paste(sprintf("%7.4f", CI_r[j]), sep=""), width=11, justify="left"))
cat("\n")
if (is.numeric(sep)) if (sep > 0) if (jj %% sep == 0) {
cat(paste(rep("-", 14 + 10 + 8 + 8 + 10 + 10 + 25), collapse="")); cat("\n")
}
jj <- jj + 1
}
cat(paste(rep("=", 14 + 10 + 8 + 8 + 10 + 10 + 25), collapse="")); cat("\n")
}
}
################################################################################
#' @title Plot Method for Local Polynomial Density Estimation and Inference
#'
#' @description The plot method for local polynomial density objects.
#'
#' @param ... Class "lpdensity" object, obtained from calling \code{\link{lpdensity}}.
#' @param alpha Numeric scalar between 0 and 1, specifies the significance level for plotting
#' confidence intervals/bands. If more than one is provided, they will be applied to each data series
#' accordingly.
#' @param type String, one of \code{"line"} (default), \code{"points"} and \code{"both"}, specifies how
#' the point estimates are plotted. If more than one is provided, they will be applied to each data series
#' accordingly.
#' @param lty Line type for point estimates, only effective if \code{type} is \code{"line"} or
#' \code{"both"}. \code{1} for solid line, \code{2} for dashed line, \code{3} for dotted line.
#' For other options, see the instructions for \code{ggplot2} or \code{\link{par}}. If
#' more than one is provided, they will be applied to each data series accordingly.
#' @param lwd Line width for point estimates, only effective if \code{type} is \code{"line"} or
#' \code{"both"}. Should be strictly positive. For other options, see the instructions for
#' \code{ggplot2} or \code{\link{par}}. If more than one is provided, they will be applied
#' to each data series accordingly.
#' @param lcol Line color for point estimates, only effective if \code{type} is \code{"line"} or
#' \code{"both"}. \code{1} for black, \code{2} for red, \code{3} for green, \code{4} for blue.
#' For other options, see the instructions for \code{ggplot2} or \code{\link{par}}. If
#' more than one is provided, they will be applied to each data series
#' accordingly.
#' @param pty Scatter plot type for point estimates, only effective if \code{type} is \code{"points"} or
#' \code{"both"}. For options, see the instructions for \code{ggplot2} or \code{\link{par}}. If
#' more than one is provided, they will be applied to each data series
#' accordingly.
#' @param pwd Scatter plot size for point estimates, only effective if \code{type} is \code{"points"} or
#' \code{"both"}. Should be strictly positive. If more than one is provided, they will be applied to each data series
#' accordingly.
#' @param pcol Scatter plot color for point estimates, only effective if \code{type} is \code{"points"} or
#' \code{"both"}. \code{1} for black, \code{2} for red, \code{3}
#' for green, \code{4} for blue.
#' For other options, see the instructions for \code{ggplot2} or \code{\link{par}}. If
#' more than one is provided, they will be applied to each data series
#' accordingly.
#' @param grid Numeric vector, specifies a subset of grid points
#' to plot point estimates. This option is effective only if \code{type} is \code{"points"} or
#' \code{"both"}; or if \code{CItype} is \code{"ebar"} or
#' \code{"all"}.
#' @param CItype String, one of \code{"region"} (shaded region, default), \code{"line"} (dashed lines),
#' \code{"ebar"} (error bars), \code{"all"} (all of the previous) or \code{"none"} (no confidence region),
#' how the confidence region should be plotted. If more than one is provided, they will be applied to each data series
#' accordingly.
#' @param CIuniform \code{TRUE} or \code{FALSE} (default), plotting either pointwise confidence intervals (\code{FALSE}) or
#' uniform confidence bands (\code{TRUE}).
#' @param CIsimul Positive integer, specifies the number of simulations used to construct critical values (default is \code{2000}). This
#' option is ignored if \code{CIuniform=FALSE}.
#' @param CIshade Numeric, specifies the opaqueness of the confidence region, should be between 0 (transparent) and
#' 1. Default is 0.2. If more than one is provided, they will be applied to each data series
#' accordingly.
#' @param CIcol Color of the confidence region. \code{1} for black, \code{2} for red, \code{3}
#' for green, \code{4} for blue.
#' For other options, see the instructions for \code{ggplot2} or \code{\link{par}}. If
#' more than one is provided, they will be applied to each data series
#' accordingly.
#' @param hist \code{TRUE} or \code{FALSE} (default), specifies whether a histogram should be added to the background.
#' @param histData Numeric vector, specifies the data used to construct the histogram plot.
#' @param histBreaks Numeric vector, specifies the breakpoints between histogram cells.
#' @param histFillCol Color of the histogram cells.
#' @param histFillShade Opaqueness of the histogram cells, should be between 0 (transparent) and
#' 1. Default is 0.2.
#' @param histLineCol Color of the histogram lines.
#' @param title,xlabel,ylabel Strings, specifies the title of the plot and labels for the x- and y-axis.
#' @param legendTitle String, specifies the legend title.
#' @param legendGroups String vector, specifies the group names used in legend.
#'
#' @return
#' A stadnard \code{ggplot} object is returned, hence can be used for further customization.
#'
#' @author
#' Matias D. Cattaneo, Princeton University. \email{cattaneo@princeton.edu}.
#'
#' Michael Jansson, University of California Berkeley. \email{mjansson@econ.berkeley.edu}.
#'
#' Xinwei Ma (maintainer), University of California San Diego. \email{x1ma@ucsd.edu}.
#'
#' @seealso \code{\link{lpdensity}} for local polynomial density estimation.
#'
#' Supported methods: \code{\link{coef.lpdensity}}, \code{\link{confint.lpdensity}},
#' \code{\link{plot.lpdensity}}, \code{\link{print.lpdensity}}, \code{\link{summary.lpdensity}},
#' \code{\link{vcov.lpdensity}}.
#'
#' @examples
#' # Generate a random sample
#' set.seed(42); X <- rnorm(2000)
#'
#' # Generate a density discontinuity at 0
#' X <- X - 0.5; X[X>0] <- X[X>0] * 2
#'
#' # Density estimation, left of 0 (scaled by the relative sample size)
#' est1 <- lpdensity(data = X[X<=0], grid = seq(-2.5, 0, 0.05), bwselect = "imse-dpi",
#' scale = sum(X<=0)/length(X))
#' # Density estimation, right of 0 (scaled by the relative sample size)
#' est2 <- lpdensity(data = X[X>0], grid = seq(0, 2, 0.05), bwselect = "imse-dpi",
#' scale = sum(X>0)/length(X))
#'
#' # Plot
#' plot(est1, est2, legendTitle="My Plot", legendGroups=c("Left", "Right"))
#'
#' # Plot uniform confidence bands
#' set.seed(42) # fix the seed for simulating critical values
#' plot(est1, est2, legendTitle="My Plot", legendGroups=c("Left", "Right"), CIuniform=TRUE)
#'
#' # Adding a histogram to the background
#' plot(est1, est2, legendTitle="My Plot", legendGroups=c("Left", "Right"),
#' hist=TRUE, histBreaks=seq(-2.4, 2, 0.2), histData=X)
#'
#' # Plot point estimates for a subset of evaluation points
#' plot(est1, est2, legendTitle="My Plot", legendGroups=c("Left", "Right"),
#' type="both", CItype="all", grid=seq(-2, 2, 0.5))
#' @export
plot.lpdensity <- function(..., alpha=NULL,
type=NULL, lty=NULL, lwd=NULL, lcol=NULL, pty=NULL, pwd=NULL, pcol=NULL, grid=NULL,
CItype=NULL, CIuniform=FALSE, CIsimul=2000, CIshade=NULL, CIcol=NULL,
hist=FALSE, histData=NULL, histBreaks=NULL, histFillCol=3, histFillShade=0.2, histLineCol="white",
title=NULL, xlabel=NULL, ylabel=NULL, legendTitle=NULL, legendGroups=NULL) {
########################################
# check how many series are passed in
########################################
x <- list(...)
nfig <- length(x)
if (nfig == 0) stop("Nothing to plot.\n")
flagToPlot <- rep(TRUE, nfig)
for (i in 1:length(x)) {
# check if is a lpbwdensity object
if (class(x[[i]])[1] == "lpbwdensity") {
flagToPlot[i] <- FALSE
warning(paste("Input ", i, " is an \"lpbwdensity\" object, which is not supported by the plot method.\n", sep=""))
next
}
# check if there is only one grid point
if (nrow(x[[i]]$Estimate) < 2) {
flagToPlot[i] <- FALSE
warning(paste("At least two grid points are needed to plot input ", i, ".\n", sep=""))
next
}
}
x <- x[flagToPlot]
nfig <- length(x)
if (nfig == 0) stop("Nothing to plot.\n")
isLpbwdensity <- FALSE
for (i in 1:nfig) {
if (class(x[[i]])[1] == "lpbwdensity") isLpbwdensity <- TRUE
}
if (isLpbwdensity) stop("The plot method does not support \"lpbwdensity\" objects.\n")
########################################
# error handling
########################################
# alpha
if (length(alpha) == 0) {
alpha <- rep(0.05, nfig)
} else if (!all(alpha>0 & alpha<1)) {
stop("Significance level incorrectly specified.\n")
} else {
alpha <- rep(alpha, length.out=nfig)
}
# plot type
if (length(type) == 0) {
type <- rep("line", nfig)
} else {
if (!all(type%in%c("line", "points", "both"))) {
stop("Plotting type incorrectly specified.\n")
}
type <- rep(type, length.out=nfig)
}
# CI type
if (length(CItype) == 0) {
CItype <- rep("region", nfig)
} else {
if (!all(CItype%in%c("region", "line", "ebar", "all", "none"))) {
stop("Confidence interval type incorrectly specified.\n")
}
CItype <- rep(CItype, length.out=nfig)
}
# line style, line width, line color
if (length(lty) == 0) {
lty <- rep(1, nfig)
} else {
lty <- rep(lty, length.out=nfig)
}
if (length(lwd) == 0) {
lwd <- rep(0.5, nfig)
} else {
lwd <- rep(lwd, length.out=nfig)
}
if (length(lcol) == 0) {
lcol <- 1:nfig
} else {
lcol <- rep(lcol, length.out=nfig)
}
# point style, point width, point color
if (length(pty) == 0) {
pty <- rep(1, nfig)
} else {
pty <- rep(pty, length.out=nfig)
}
if (length(pwd) == 0) {
pwd <- rep(1, nfig)
} else {
pwd <- rep(pwd, length.out=nfig)
}
if (length(pcol) == 0) {
pcol <- lcol
} else {
pcol <- rep(pcol, length.out=nfig)
}
# CI shade, CI color
if (length(CIshade) == 0) {
CIshade <- rep(0.2, nfig)
} else {
CIshade <- rep(CIshade, length.out=nfig)
}
if (length(CIcol) == 0) {
CIcol <- lcol
} else {
CIcol <- rep(CIcol, length.out=nfig)
}
# legend
# New in v0.2.1 to handle legend
if (length(legendTitle) == 0) {
legendTitle <- ""
} else {
legendTitle <- legendTitle[1]
}
if (length(legendGroups) > 0) {
legendGroups <- rep(legendGroups, length.out=nfig)
legend_default <- FALSE
} else {
legend_default <- TRUE
}
# grid
if (!is.null(grid)) {
if (!is.numeric(grid)) {
stop("Option grid incorrectly specified.\n")
}
}
########################################
# initializing plot
########################################
if (hist & !is.null(histData)) {
histData <- as.data.frame(histData)
colnames(histData) <- c("v1")
if(is.null(histBreaks)) { histBreaks <- seq(from=min(histData[, 1]), to=max(histData[, 1]), length.out=21) }
histScale <- mean(histData[, 1] >= min(histBreaks) & histData[, 1] <= max(histBreaks))
temp_plot <- ggplot() +
geom_histogram(data=histData, aes(x=v1, y=after_stat(density)*histScale), breaks=histBreaks, fill=histFillCol, col=histLineCol, alpha=histFillShade) +
theme_bw() #+ theme(legend.position="none")
} else {
temp_plot <- ggplot() + theme_bw() #+ theme(legend.position="none")
}
#CI_l <- CI_r <- f_p <- Sname <- v1 <- ..density.. <- NULL
CI_l <- CI_r <- f_p <- Sname <- v1 <- NULL
########################################
# looping over input models
########################################
### all colors
col_all <- lty_all <- pty_all <- v_all <- c()
estRangeL <- estRangeR <- c() # estimation range
for (i in 1:nfig) {
# get derivative order
v_all <- c(v_all, x[[i]]$opt$v)
# get ploting indices
if (is.null(grid)) {
plotIndex <- 1:nrow(x[[i]]$Estimate)
} else {
gridTemp <- grid[grid >= min(x[[i]]$Estimate[, "grid"]) & grid <= max(x[[i]]$Estimate[, "grid"])]
if (length(gridTemp) == 0) {
plotIndex <- NULL
} else {
plotIndex <- rep(NA, length(gridTemp))
for (ii in 1:length(gridTemp)) {
plotIndex[ii] <- which.min(abs(gridTemp[ii]-x[[i]]$Estimate[, "grid"]))
}
plotIndex <- unique(plotIndex)
}
}
estRangeL <- min(estRangeL, min(x[[i]]$Estimate[, "grid"]))
estRangeR <- max(estRangeR, max(x[[i]]$Estimate[, "grid"]))
data_x <- data.frame(x[[i]]$Estimate[, c("grid", "f_p", "f_q", "se_p", "se_q"), drop=FALSE])
if (x[[i]]$opt$q == x[[i]]$opt$p) {
data_x$f_q <- data_x$f_p; data_x$se_q <- data_x$se_p
}
# critical value
if (CIuniform) {
if (length(CIsimul) == 0) { CIsimul <- 2000 }
if (!is.numeric(CIsimul) | is.na(CIsimul)) {
warning("Option CIsimul incorrectly specified. Will only plot pointwise confidence intervals.\n")
z_val <- qnorm(1 - alpha[i]/2)
} else if (ceiling(CIsimul)<2) {
warning("Option CIsimul incorrectly specified. Will only plot pointwise confidence intervals.\n")
z_val <- qnorm(1 - alpha[i]/2)
} else {
CIsimul <- ceiling(CIsimul)
if (x[[i]]$opt$q == x[[i]]$opt$p) {
x[[i]]$CovMat_q <- x[[i]]$CovMat_p
}
corrMat <- sweep(sweep(x[[i]]$CovMat_q, MARGIN=1, FUN="*", STATS=1/data_x$se_q), MARGIN=2, FUN="*", STATS=1/data_x$se_q)
normalSimu <- try(
mvrnorm(n=CIsimul, mu=rep(0,nrow(corrMat)), Sigma=corrMat),
silent=TRUE)
if (is.character(normalSimu)) {
print(normalSimu)
warning("Variance-Covariance is not positive semidefinite. Will only plot pointwise confidence intervals.\n")
z_val <- qnorm(1 - alpha[i]/2)
} else {
z_val <- quantile(apply(normalSimu, MARGIN=1, FUN=function(x) {max(abs(x))}), 1 - alpha[i])
}
}
} else {
z_val <- qnorm(1 - alpha[i]/2)
}
data_x$CI_l <- data_x$f_q - z_val * data_x$se_q
data_x$CI_r <- data_x$f_q + z_val * data_x$se_q
# New in v0.2.1 to handle legend
if (legend_default) {
data_x$Sname <- paste("Series", i, sep=" ")
legendGroups <- c(legendGroups, data_x$Sname)
} else {
data_x$Sname <- legendGroups[i]
}
########################################
# add CI regions to the plot
if (CItype[i]%in%c("region", "all"))
temp_plot <- temp_plot + geom_ribbon(data=data_x, aes(x=grid, ymin=CI_l, ymax=CI_r), alpha=CIshade[i], fill=CIcol[i])
########################################
# add CI lines to the plot
if (CItype[i]%in%c("line", "all"))
temp_plot <- temp_plot + geom_line(data=data_x, aes(x=grid, y=CI_l), linetype=2, alpha=1, col=CIcol[i]) +
geom_line(data=data_x, aes(x=grid, y=CI_r), linetype=2, alpha=1, col=CIcol[i])
#if (CItype[i]%in%c("line", "all"))
# temp_plot <- temp_plot + geom_line(data=data_x, aes(x=grid, y=CI_l), linetype=2, alpha=CIshade[i], col=CIcol[i]) +
# geom_line(data=data_x, aes(x=grid, y=CI_r), linetype=2, alpha=CIshade[i], col=CIcol[i])
########################################
# add error bars to the plot
if (CItype[i]%in%c("ebar", "all") & !is.null(plotIndex))
temp_plot <- temp_plot + geom_errorbar(data=data_x[plotIndex, ], aes(x=grid, ymin=CI_l, ymax=CI_r), alpha=1, col=CIcol[i], linetype=1)
#if (CItype[i]%in%c("ebar", "all") & !is.null(plotIndex))
# temp_plot <- temp_plot + geom_errorbar(data=data_x[plotIndex, ], aes(x=grid, ymin=CI_l, ymax=CI_r), alpha=CIshade[i], col=CIcol[i], linetype=1)
########################################
# add lines to the plot
if (type[i]%in%c("line", "both")) {
temp_plot <- temp_plot + geom_line(data=data_x, aes(x=grid, y=f_p, colour=Sname, linetype=Sname), size=lwd[i])
}
########################################
# add points to the plot
if (type[i]%in%c("points", "both") & !is.null(plotIndex)) {
temp_plot <- temp_plot + geom_point(data=data_x[plotIndex, ], aes(x=grid, y=f_p, colour=Sname, shape=Sname), size=pwd[i])
}
if (type[i] == "line") {
col_all <- c(col_all, lcol[i])
lty_all <- c(lty_all, lty[i])
pty_all <- c(pty_all, NA)
} else if (type[i] == "both") {
col_all <- c(col_all, lcol[i])
lty_all <- c(lty_all, lty[i])
pty_all <- c(pty_all, pty[i])
} else {
col_all <- c(col_all, pcol[i])
lty_all <- c(lty_all, NA)
pty_all <- c(pty_all, pty[i])
}
}
########################################
# change color, line type and point shape back, and customize legend
########################################
# New in v0.2.1 to handle legend
index <- sort.int(legendGroups, index.return=TRUE)$ix
temp_plot <- temp_plot + scale_color_manual(values = col_all[index]) +
scale_linetype_manual(values = lty_all[index]) +
scale_shape_manual(values = pty_all[index]) +
guides(colour=guide_legend(title=legendTitle)) +
guides(linetype=guide_legend(title=legendTitle)) +
guides(shape=guide_legend(title=legendTitle))
########################################
# add title, x and y labs
########################################
if (is.null(ylabel)) {
if (all(v_all == v_all[1])) {
if (v_all[1] == 0) {
ylabel <- "Distribution function"
} else if (v_all[1] == 1) {
ylabel <- "Density"
} else {
ylabel <- paste("Density derivative (v=", v_all[1], ")", sep="")
}
} else {
ylabel <- ""
}
}
if (is.null(xlabel)) {
xlabel <- ""
}
if (is.null(title)) {
title <- ""
}
temp_plot <- temp_plot + labs(x=xlabel, y=ylabel) + ggtitle(title)
# check plotting range vs estimation range
if (!is.null(grid)) {
if (min(grid) < estRangeL | max(grid) > estRangeR) {
warning("The plotting range exceeds the original estimation range. Option plot(..., grid=) should be within the estimation range specified by lpdensity(..., grid=).\n")
}
}
########################################
# return the plot
########################################
return (temp_plot)
}
################################################################################
#' @title Plot Method for Local Polynomial Density Estimation and Inference
#'
#' @description This has been replaced by \code{\link{plot.lpdensity}}.
#'
#' @param ... Class "lpdensity" object, obtained from calling \code{\link{lpdensity}}.
#' @param alpha Numeric scalar between 0 and 1, specifies the significance level for plotting
#' confidence intervals/bands. If more than one is provided, they will be applied to each data series
#' accordingly.
#' @param type String, one of \code{"line"} (default), \code{"points"} and \code{"both"}, specifies how
#' the point estimates are plotted. If more than one is provided, they will be applied to each data series
#' accordingly.
#' @param lty Line type for point estimates, only effective if \code{type} is \code{"line"} or
#' \code{"both"}. \code{1} for solid line, \code{2} for dashed line, \code{3} for dotted line.
#' For other options, see the instructions for \code{ggplot2} or \code{\link{par}}. If
#' more than one is provided, they will be applied to each data series accordingly.
#' @param lwd Line width for point estimates, only effective if \code{type} is \code{"line"} or
#' \code{"both"}. Should be strictly positive. For other options, see the instructions for
#' \code{ggplot2} or \code{\link{par}}. If more than one is provided, they will be applied
#' to each data series accordingly.
#' @param lcol Line color for point estimates, only effective if \code{type} is \code{"line"} or
#' \code{"both"}. \code{1} for black, \code{2} for red, \code{3} for green, \code{4} for blue.
#' For other options, see the instructions for \code{ggplot2} or \code{\link{par}}. If
#' more than one is provided, they will be applied to each data series
#' accordingly.
#' @param pty Scatter plot type for point estimates, only effective if \code{type} is \code{"points"} or
#' \code{"both"}. For options, see the instructions for \code{ggplot2} or \code{\link{par}}. If
#' more than one is provided, they will be applied to each data series
#' accordingly.
#' @param pwd Scatter plot size for point estimates, only effective if \code{type} is \code{"points"} or
#' \code{"both"}. Should be strictly positive. If more than one is provided, they will be applied to each data series
#' accordingly.
#' @param pcol Scatter plot color for point estimates, only effective if \code{type} is \code{"points"} or
#' \code{"both"}. \code{1} for black, \code{2} for red, \code{3}
#' for green, \code{4} for blue.
#' For other options, see the instructions for \code{ggplot2} or \code{\link{par}}. If
#' more than one is provided, they will be applied to each data series
#' accordingly.
#' @param grid Numeric vector, specifies a subset of grid points
#' to plot point estimates. This option is effective only if \code{type} is \code{"points"} or
#' \code{"both"}; or if \code{CItype} is \code{"ebar"} or
#' \code{"all"}.
#' @param CItype String, one of \code{"region"} (shaded region, default), \code{"line"} (dashed lines),
#' \code{"ebar"} (error bars), \code{"all"} (all of the previous) or \code{"none"} (no confidence region),
#' how the confidence region should be plotted. If more than one is provided, they will be applied to each data series
#' accordingly.
#' @param CIuniform \code{TRUE} or \code{FALSE} (default), plotting either pointwise confidence intervals (\code{FALSE}) or
#' uniform confidence bands (\code{TRUE}).
#' @param CIsimul Positive integer, specifies the number of simulations used to construct critical values (default is \code{2000}). This
#' option is ignored if \code{CIuniform=FALSE}.
#' @param CIshade Numeric, specifies the opaqueness of the confidence region, should be between 0 (transparent) and
#' 1. Default is 0.2. If more than one is provided, they will be applied to each data series
#' accordingly.
#' @param CIcol Color of the confidence region. \code{1} for black, \code{2} for red, \code{3}
#' for green, \code{4} for blue.
#' For other options, see the instructions for \code{ggplot2} or \code{\link{par}}. If
#' more than one is provided, they will be applied to each data series
#' accordingly.
#' @param hist \code{TRUE} or \code{FALSE} (default), specifies whether a histogram should be added to the background.
#' @param histData Numeric vector, specifies the data used to construct the histogram plot.
#' @param histBreaks Numeric vector, specifies the breakpoints between histogram cells.
#' @param histFillCol Color of the histogram cells.
#' @param histFillShade Opaqueness of the histogram cells, should be between 0 (transparent) and
#' 1. Default is 0.2.
#' @param histLineCol Color of the histogram lines.
#' @param title,xlabel,ylabel Strings, specifies the title of the plot and labels for the x- and y-axis.
#' @param legendTitle String, specifies the legend title.
#' @param legendGroups String vector, specifies the group names used in legend.
#'
#' @return
#' A stadnard \code{ggplot} object is returned, hence can be used for further customization.
#'
#' @author
#' Matias D. Cattaneo, Princeton University. \email{cattaneo@princeton.edu}.
#'
#' Michael Jansson, University of California Berkeley. \email{mjansson@econ.berkeley.edu}.
#'
#' Xinwei Ma (maintainer), University of California San Diego. \email{x1ma@ucsd.edu}.
#'
#' @export
lpdensity.plot <- plot.lpdensity
################################################################################
#' Coef Method for Local Polynomial Density Estimation and Inference
#'
#' @description The coef method for local polynomial density objects.
#'
#' @param object Class "lpdensity" object, obtained by calling \code{\link{lpdensity}}.
#' @param ... Additional options.
#'
#' @return
#' A matrix containing grid points and density estimates using p- and q-th order local polynomials.
#'
#' @author
#' Matias D. Cattaneo, Princeton University. \email{cattaneo@princeton.edu}.
#'
#' Michael Jansson, University of California Berkeley. \email{mjansson@econ.berkeley.edu}.
#'
#' Xinwei Ma (maintainer), University of California San Diego. \email{x1ma@ucsd.edu}.
#'
#' @seealso \code{\link{lpdensity}} for local polynomial density estimation.
#'
#' Supported methods: \code{\link{coef.lpdensity}}, \code{\link{confint.lpdensity}},
#' \code{\link{plot.lpdensity}}, \code{\link{print.lpdensity}}, \code{\link{summary.lpdensity}},
#' \code{\link{vcov.lpdensity}}.
#'
#' @examples
#' # Generate a random sample
#' set.seed(42); X <- rnorm(2000)
#'
#' # Estimate density and report results
#' coef(lpdensity(data = X, bwselect = "imse-dpi"))
#'
#' @export
coef.lpdensity <- function(object, ...) {
object$Estimate[, c("grid", "f_p", "f_q")]
}
################################################################################
#' Vcov Method for Local Polynomial Density Estimation and Inference
#'
#' @description The vcov method for local polynomial density objects.
#'
#' @param object Class "lpdensity" object, obtained by calling \code{\link{lpdensity}}.
#' @param ... Additional options.
#'
#' @return
#' \item{stdErr}{A matrix containing grid points and standard errors using p- and q-th order local polynomials.}
#' \item{CovMat_p}{The variance-covariance matrix corresponding to \code{f_p}.}
#' \item{CovMat_q}{The variance-covariance matrix corresponding to \code{f_q}.}
#'
#' @author
#' Matias D. Cattaneo, Princeton University. \email{cattaneo@princeton.edu}.
#'
#' Michael Jansson, University of California Berkeley. \email{mjansson@econ.berkeley.edu}.
#'
#' Xinwei Ma (maintainer), University of California San Diego. \email{x1ma@ucsd.edu}.
#'
#' @seealso \code{\link{lpdensity}} for local polynomial density estimation.
#'
#' Supported methods: \code{\link{coef.lpdensity}}, \code{\link{confint.lpdensity}},
#' \code{\link{plot.lpdensity}}, \code{\link{print.lpdensity}}, \code{\link{summary.lpdensity}},
#' \code{\link{vcov.lpdensity}}.
#'
#' @examples
#' # Generate a random sample
#' set.seed(42); X <- rnorm(2000)
#'
#' # Estimate density and report results
#' vcov(lpdensity(data = X, bwselect = "imse-dpi"))
#'
#' @export
vcov.lpdensity <- function(object, ...) {
if (class(object)[1] == "lpbwdensity") stop("The vcov method does not support \"lpbwdensity\" objects.\n")
return(list(stdErr=object$Estimate[, c("grid", "se_p", "se_q")], CovMat_p=object$CovMat_p, CovMat_q=object$CovMat_q))
}
################################################################################
#' Confint Method for Local Polynomial Density Estimation and Inference
#'
#' @description The confint method for local polynomial density objects.
#'
#' @param object Class "lpdensity" object, obtained by calling \code{\link{lpdensity}}.
#' @param parm Integer, indicating which parameters are to be given confidence intervals.
#' @param level Numeric scalar between 0 and 1, the significance level for computing confidence intervals
#' @param ... Additional options, including (i) \code{grid} specifies a subset of grid points
#' to display the bandwidth; (ii) \code{gridIndex} specifies the indices of grid points
#' to display the bandwidth (this is the same as \code{parm}); (iii) \code{alpha} specifies the significance level
#' (this is \code{1-level}); (iv)
#' \code{CIuniform} specifies whether displaying pointwise confidence intervals (\code{FALSE}, default) or
#' the uniform confidence band (\code{TRUE}); (v) \code{CIsimul} specifies the number of simulations used
#' to construct critical values (default is 2000).
#'
#' @return
#' A matrix containing grid points and confidence interval end points using p- and q-th order local polynomials.
#'
#' @author
#' Matias D. Cattaneo, Princeton University. \email{cattaneo@princeton.edu}.
#'
#' Michael Jansson, University of California Berkeley. \email{mjansson@econ.berkeley.edu}.
#'
#' Xinwei Ma (maintainer), University of California San Diego. \email{x1ma@ucsd.edu}.
#'
#' @seealso \code{\link{lpdensity}} for local polynomial density estimation.
#'
#' Supported methods: \code{\link{coef.lpdensity}}, \code{\link{confint.lpdensity}},
#' \code{\link{plot.lpdensity}}, \code{\link{print.lpdensity}}, \code{\link{summary.lpdensity}},
#' \code{\link{vcov.lpdensity}}.
#'
#' @examples
#' # Generate a random sample
#' set.seed(42); X <- rnorm(2000)
#'
#' # Estimate density and report 95% confidence intervals
#' est1 <- lpdensity(data = X, bwselect = "imse-dpi")
#' confint(est1)
#'
#' # Report results for a subset of grid points
#' confint(est1, parm=est1$Estimate[4:10, "grid"])
#' confint(est1, grid=est1$Estimate[4:10, "grid"])
#' confint(est1, gridIndex=4:10)
#'
#' # Report the 99% uniform confidence band
#' # Fix the seed for simulating critical values
#' set.seed(42); confint(est1, level=0.99, CIuniform=TRUE)
#' set.seed(42); confint(est1, alpha=0.01, CIuniform=TRUE)
#'
#' @export
confint.lpdensity <- function(object, parm = NULL, level = NULL, ...) {
x <- object
if (class(x)[1] == "lpbwdensity") stop("The confint method does not support \"lpbwdensity\" objects.\n")
args <- list(...)
if (!is.null(parm)) { args[['grid']] <- parm }
if (!is.null(level)) { args[['alpha']] <- 1 - level }
if (is.null(args[['alpha']])) { alpha <- 0.05 } else { alpha <- args[['alpha']] }
if (is.null(args[['CIuniform']])) { CIuniform <- FALSE } else { CIuniform <- args[['CIuniform']] }
if (is.null(args[['CIsimul']])) { CIsimul <- 2000 } else { sep <- args[['CIsimul']] }
if (is.null(args[['grid']]) & is.null(args[['gridIndex']])) {
gridIndex <- 1:nrow(x$Estimate)
} else if (is.null(args[['grid']]) & !is.null(args[['gridIndex']])) {
gridIndex <- args[['gridIndex']]
if (is.null(gridIndex)) {
gridIndex <- 1:nrow(x$Estimate)
} else if (!all(gridIndex %in% 1:nrow(x$Estimate))) {
stop("Option gridIndex incorrectly specified.\n")
}
} else {
grid <- args[['grid']]
if (is.null(grid)) {
gridIndex <- 1:nrow(x$Estimate)
} else if (!is.numeric(grid)) {
stop("Option param/grid incorrectly specified.\n")
} else {
gridIndex <- rep(NA, length(grid))
for (j in 1:length(grid)) {
gridIndex[j] <- which.min(abs(x$Estimate[, "grid"]-grid[j]))
#gridIndex <- unique(gridIndex)
}
}
}
Estimate <- matrix(NA, nrow=length(gridIndex), ncol=5)
Estimate[, 1] <- x$Estimate[gridIndex, "grid"]
if (CIuniform) {
if (length(CIsimul) == 0) { CIsimul <- 2000 }
if (!is.numeric(CIsimul) | is.na(CIsimul)) {
warning("Option CIsimul incorrectly specified. Will only plot pointwise confidence intervals.\n")
CIuniform <- FALSE
z_val <- qnorm(1 - alpha/2)
} else if (ceiling(CIsimul)<2) {
warning("Option CIsimul incorrectly specified. Will only plot pointwise confidence intervals.\n")
CIuniform <- FALSE
z_val <- qnorm(1 - alpha/2)
} else {
CIsimul <- ceiling(CIsimul)
if (x$opt$q == x$opt$p) {
x$CovMat_q <- x$CovMat_p
x$Estimate[, "se_q"] <- x$Estimate[, "se_p"]
}
corrMat <- sweep(sweep(x$CovMat_q, MARGIN=1, FUN="*", STATS=1/x$Estimate[, "se_q"]), MARGIN=2, FUN="*", STATS=1/x$Estimate[, "se_q"])
normalSimu <- try(
mvrnorm(n=CIsimul, mu=rep(0,nrow(corrMat)), Sigma=corrMat),
silent=TRUE)
if (is.character(normalSimu)) {
print(normalSimu)
warning("Variance-Covariance is not positive semidefinite. Will only plot pointwise confidence intervals.\n")
CIuniform <- FALSE
z_val <- qnorm(1 - alpha/2)
} else {
z_val <- quantile(apply(normalSimu, MARGIN=1, FUN=function(x) {max(abs(x))}), 1 - alpha)
}
}
} else {
z_val <- qnorm(1 - alpha/2)
}
Estimate[, 2] <- x$Estimate[gridIndex, "f_p"] - z_val * x$Estimate[gridIndex, "se_p"]
Estimate[, 3] <- x$Estimate[gridIndex, "f_p"] + z_val * x$Estimate[gridIndex, "se_p"]
Estimate[, 4] <- x$Estimate[gridIndex, "f_q"] - z_val * x$Estimate[gridIndex, "se_q"]
Estimate[, 5] <- x$Estimate[gridIndex, "f_q"] + z_val * x$Estimate[gridIndex, "se_q"]
colnames(Estimate) <- c("grid", "CI_l_p", "CI_r_p", "CI_l_q", "CI_r_q")
#if (x$opt$p == x$opt$q) Estimate <- Estimate[, 1:3]
return(Estimate)
}
Try the lpdensity package in your browser
Any scripts or data that you put into this service are public.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.