investr (version 1.1.0)

plotFit: Plotting Confidence/Prediction Bands

Description

Plots fitted model for an object of class lm or nlsvwith the option of adding a confidence and/or prediction band.

Usage

plotFit(object, ...)
## S3 method for class 'lm':
plotFit(object, interval = c("none", "both", "confidence",
  "prediction"), level = 0.95, adjust = c("none", "Bonferroni", "Scheffe"),
  k, ..., shade = FALSE, extend.range = FALSE, hide = TRUE,
  col.conf = if (shade) grey(0.7) else "black", col.pred = if (shade)
  grey(0.9) else "black", border.conf = col.conf, border.pred = col.pred,
  col.fit = "black", lty.conf = if (shade) 1 else 2, lty.pred = if (shade)
  1 else 3, lty.fit = 1, lwd.conf = 1, lwd.pred = 1, lwd.fit = 1,
  n = 500, xlab, ylab, xlim, ylim = NULL)
## S3 method for class 'nls':
plotFit(object, interval = c("none", "both", "confidence",
  "prediction"), level = 0.95, adjust = c("none", "Bonferroni", "Scheffe"),
  k, ..., shade = FALSE, extend.range = FALSE, hide = TRUE,
  col.conf = if (shade) grey(0.7) else "black", col.pred = if (shade)
  grey(0.9) else "black", border.conf = col.conf, border.pred = col.pred,
  col.fit = "black", lty.conf = if (shade) 1 else 2, lty.pred = if (shade)
  1 else 3, lty.fit = 1, lwd.conf = 1, lwd.pred = 1, lwd.fit = 1,
  n = 500, xlab, ylab, xlim, ylim = NULL)

Arguments

object

An object that inherits from class lm or nls.

interval

A character string indicating if a prediction band, confidence band, both, or none should be plotted.

level

The desired confidence level.

adjust

A character string indicating the type of adjustment (if any) to make to the confidence/prediction bands.

k

An integer to be used in computing the critical value for the confidence/prediction bands. Only needed when adjust = "Bonferroni" or when adjust = "Scheffe" and interval = "prediction".

shade

A logical value indicating if the band should be shaded.

extend.range

A logical value indicating if the fitted regression line and bands (if any) should extend to the edges of the plot. Default is FALSE.

col.conf

Shade color for confidence band.

col.pred

Shade color for prediction band.

col.fit

The color to use for the fitted line.

border.conf

The color to use for the confidence band border.

border.pred

The color to use for the prediction band border.

lty.conf

Line type to use for confidence band border.

lty.pred

Line type to use for prediction band border.

lty.fit

Line type to use for the fitted regression line.

lwd.conf

Line width to use for confidence band border.

lwd.pred

Line width to use for prediction band border.

lwd.fit

Line width to use for the fitted regression line.

n

The number of predictor values at which to evaluate the fitted model (larger implies a smoother plot).

xlab

A title for the x axis.

ylab

A title for the y axis.

xlim

The x limits (x1, x2) of the plot.

ylim

The y limits (y1, y2) of the plot.

hide

A logical value indicating if the fitted model should be plotted on top of the points (FALSE) or behind them (TRUE). Default is TRUE.

...

Additional optional arguments passed on to plot.

References

Bates, D. M., and Watts, D. G. Nonlinear Regression Analysis and its Applications. New York: Wiley, 2007.

F. Baty and M. L. Delignette-Muller (2012), nlstools: Tools for Nonlinear Regression Diagnostics.

Examples

## A linear regression example
data(cars, package = "datasets")
library(splines)
cars.lm1 <- lm(dist ~ speed, data = cars)
cars.lm2 <- lm(dist ~ speed + I(speed^2), data = cars)
cars.lm3 <- lm(dist ~ speed + I(speed^2) + I(speed^3), data = cars)
cars.lm4 <- lm(dist ~ ns(speed, df = 3), data = cars)
par(mfrow = c(2, 2))
plotFit(cars.lm1, interval = "both", xlim = c(-10, 40), ylim = c(-50, 150),
        main = "linear")
plotFit(cars.lm2, interval = "both", xlim = c(-10, 40), ylim = c(-50, 150),
        main = "quadratic")
plotFit(cars.lm3, interval = "both", xlim = c(-10, 40), ylim = c(-50, 150),
        main = "cubic")
plotFit(cars.lm4, interval = "both", xlim = c(-10, 40), ylim = c(-50, 150),
        main = "cubic spline")

## A nonlinear regression example
par(mfrow = c(1, 1))
library(RColorBrewer) # requires that RColorBrewer be installed
blues <- brewer.pal(9, "Blues")
data(Puromycin, package = "datasets")
Puromycin2 <- Puromycin[Puromycin$state == "treated", ][, 1:2]
Puro.nls <- nls(rate ~ Vm * conc/(K + conc), data = Puromycin2,
                start = c(Vm = 200, K = 0.05))
plotFit(Puro.nls, interval = "both", pch = 19, shade = TRUE,
        col.conf = blues[4], col.pred = blues[2])