qqdiag: qqdiag generates a qqplot with a histogram of residuals

In haddonm/r4cpue: Functions to assist with CPUE standardization

Description

qqdiag generates a qqplot with a complementary histogram of the residuals to illustrate the proportion of all residuals along the qqline. If the qqline deviates from the expected straigt line, which is red i colour to make for simpler comparisons, then the histogram enables one to estiamte what proportion of records deviate from normality. The zero point is identified with a line, as are the approximate 5 below the blue lines, with < 90 proportions in each class. To get a more precise estimate use the invisibly returned histogram values.

Usage

qqdiag(
  inmodel,
  plotrug = FALSE,
  bins = NA,
  hline = 0,
  xinc = 100,
  yinc = 1,
  ylab = "residuals"
)

Arguments

inmodel	the optimum model being considered
plotrug	a logical term determining whether a rug is plotted on the qqplot.
bins	defaults to NA, but can be set to a given series
hline	Include some horizontal lines on the histogram. defaults to 0.
xinc	the increment for tick marks on the xaxis of the histogram
yinc	the increment for tick marks on the y-axis of the histogram
ylab	the y-axis label for the histogram, defaults to 'residuals'

Value

plots a graph and invisibly returns the output from the histogram

Examples

y <- rep(1:100,2)
x <- rnorm(200,mean=10,sd=1)
model <- lm(y ~ x)
dev.new(width=6,height=3.5,noRStudioGD = TRUE)
par(mai=c(0.45,0.45,0.15,0.05),font.axis=7)
qqdiag(model,xinc=1,yinc=10,bins=seq(-55,50,2.5))

haddonm/r4cpue documentation built on May 11, 2020, 1:31 a.m.