qqnormSim {StMoSim}

Quantile-Quantile plot with several Gaussian simulations.

Description

Plots a QQ plot of the variable x with nSim Gaussian simulations.

Usage

qqnormSim(x, nSim = 500, mOfVar = "mad",
  main = "Normal Q-Q Plot - SIM", xlab = "Theoretical Quantiles",
  ylab = "Sample Quantiles", qqnormCol = "black", qqnormPch = 1,
  qqlineCol = "#cdd2d015", qqlineLwd = 3)

## S4 method for signature 'lm'
qqnormSim(x, nSim = 500, mOfVar = "mad",
  main = "Normal Q-Q Plot - SIM", xlab = "Theoretical Quantiles",
  ylab = "Sample Quantiles", qqnormCol = "black", qqnormPch = 1,
  qqlineCol = "#cdd2d015", qqlineLwd = 3)

## S4 method for signature 'numeric'
qqnormSim(x, nSim = 500, mOfVar = "mad",
  main = "Normal Q-Q Plot - SIM", xlab = "Theoretical Quantiles",
  ylab = "Sample Quantiles", qqnormCol = "black", qqnormPch = 1,
  qqlineCol = "#cdd2d015", qqlineLwd = 3)

Arguments

x

a lm-object or a numeric vector. If it's a lm-object its residuals are plotted.

nSim

[optional] the number of simulations you like to add to the plot.

mOfVar

[optinal] a measure of variation. ("mad" or "sd")

main

[optional] an overall title for the plot.

xlab

[optional] a title for the x axis.

ylab

[optional] a title for the y axis.

qqnormCol

[optional] color of the obervations in the plot.

qqnormPch

[optional] point character of the observations in the plot.

qqlineCol

[optional] color of the simulations in the plot.

qqlineLwd

[optional] line width of the simulations. should not be higher than 3.

Details

Two estimators are required for the simulation of the normal distribution. Since the normal distribution is a two-parameter family distribution. Default measure of location is the mean. Default measure of variation is the mad. This gives a robust estimation of the standard deviation even if there are outliers in the sample. Likewise this can be changed with the parameter mOfVar.

Value

invisible(NULL)

Author(s)

Matthias Salvisberg <[e-mail]>

See Also

the basic graph corresponds to qqnorm

Examples

## Not run: 

######## qqnorm vs. qqnormSim ########

par(mfrow = c(1,2))
x<- rnorm(100)
qqnorm(x)
qqline(x)
qqnormSim(x)
par(mfrow = c(1,1))

######## basic functionality/arguments ########

# The observations should behave like a simulation, 
# because the observations are sampled from a Gaussian distribution.
qqnormSim(x = rnorm(100))

# If you don't feel comfortable with the mad as 
# measure of variation you can change it to the standard deviation.
qqnormSim(x = rnorm(100),
          mOfVar = "sd")

# On the first glance its obvious that this sample 
# doesn't originate from a Gaussian distribution due to the heavy tails.
qqnormSim(x = rt(100,df = 4))

Reduce the simulation tracks from 500 to 50. (500 is default).
Not recommended unless you have not enough computation power.
qqnormSim(x = rnorm(100), 
          nSim = 50)

######## graphical arguments ########

# set title and axes labels.
qqnormSim(x = rnorm(100), 
          main = "main title",
          xlab = "x-axis label",
          ylab = "y-axis label")
          
# I don't recommend fancy colors, unless you need it for your corporate identity.
qqnormSim(x = rnorm(100), 
          qqnormCol = "#ff0000",
          qqnormPch = 16,
          qqlineCol = "greenyellow",
          qqlineLwd = 1)


## End(Not run)