| % File src/library/stats/man/predict.loess.Rd |
| % Part of the R package, https://www.R-project.org |
| % Copyright 1995-2020 R Core Team |
| % Distributed under GPL 2 or later |
| |
| \name{predict.loess} |
| \alias{predict.loess} |
| \title{Predict Loess Curve or Surface} |
| \description{ |
| Predictions from a \code{loess} fit, optionally with standard errors. |
| } |
| \usage{ |
| \method{predict}{loess}(object, newdata = NULL, se = FALSE, |
| na.action = na.pass, \dots) |
| } |
| \arguments{ |
| \item{object}{an object fitted by \code{loess}.} |
| \item{newdata}{an optional data frame in which to look for variables with |
| which to predict, or a matrix or vector containing exactly the variables |
| needs for prediction. If missing, the original data points are used.} |
| \item{se}{should standard errors be computed?} |
| \item{na.action}{function determining what should be done with missing |
| values in data frame \code{newdata}. The default is to predict \code{NA}.} |
| \item{\dots}{arguments passed to or from other methods.} |
| } |
| \details{ |
| The standard errors calculation \code{se = TRUE} is slower than |
| prediction, notably as it needs a relatively large workspace (memory), |
| notably matrices of dimension \eqn{N \times Nf}{N * Nf} where \eqn{f = |
| }\code{span}, i.e., \code{se = TRUE} is \eqn{O(N^2)} |
| and hence stops when the sample size \eqn{N} is larger than about 40'600 |
| (for default \code{span = 0.75}). |
| |
| When the fit was made using \code{surface = "interpolate"} (the |
| default), \code{predict.loess} will not extrapolate -- so points outside |
| an axis-aligned hypercube enclosing the original data will have |
| missing (\code{NA}) predictions and standard errors. |
| } |
| \value{ |
| If \code{se = FALSE}, a vector giving the prediction for each row of |
| \code{newdata} (or the original data). If \code{se = TRUE}, a list |
| containing components |
| \item{fit}{the predicted values.} |
| \item{se}{an estimated standard error for each predicted value.} |
| \item{residual.scale}{the estimated scale of the residuals used in |
| computing the standard errors.} |
| \item{df}{an estimate of the effective degrees of freedom used in |
| estimating the residual scale, intended for use with t-based |
| confidence intervals. } |
| If \code{newdata} was the result of a call to |
| \code{\link{expand.grid}}, the predictions (and s.e.'s if requested) |
| will be an array of the appropriate dimensions. |
| |
| Predictions from infinite inputs will be \code{NA} since \code{loess} |
| does not support extrapolation. |
| } |
| \author{ |
| B. D. Ripley, based on the \code{cloess} package of Cleveland, |
| Grosse and Shyu. |
| } |
| \note{ |
| Variables are first looked for in \code{newdata} and then searched for |
| in the usual way (which will include the environment of the formula |
| used in the fit). A warning will be given if the |
| variables found are not of the same length as those in \code{newdata} |
| if it was supplied. |
| } |
| |
| \seealso{\code{\link{loess}}} |
| |
| \examples{ |
| cars.lo <- loess(dist ~ speed, cars) |
| predict(cars.lo, data.frame(speed = seq(5, 30, 1)), se = TRUE) |
| # to get extrapolation |
| cars.lo2 <- loess(dist ~ speed, cars, |
| control = loess.control(surface = "direct")) |
| predict(cars.lo2, data.frame(speed = seq(5, 30, 1)), se = TRUE) |
| } |
| \keyword{smooth} |