residual_plots function

Several types of residual plots

Residual plots for a output model of class performs_ammi, waas, anova_ind, and anova_joint. Seven types of plots are produced: (1) Residuals vs fitted, (2) normal Q-Q plot for the residuals, (3) scale-location plot (standardized residuals vs Fitted Values), (4) standardized residuals vs Factor-levels, (5) Histogram of raw residuals and (6) standardized residuals vs observation order, and (7) 1:1 line plot

residual_plots(
  x,
  var = 1,
  conf = 0.95,
  labels = FALSE,
  plot_theme = theme_metan(),
  band.alpha = 0.2,
  point.alpha = 0.8,
  fill.hist = "gray",
  col.hist = "black",
  col.point = "black",
  col.line = "red",
  col.lab.out = "red",
  size.lab.out = 2.5,
  size.tex.lab = 10,
  size.shape = 1.5,
  bins = 30,
  which = c(1:4),
  ncol = NULL,
  nrow = NULL,
  ...
)

Arguments

• x: An object of class performs_ammi, waas, anova_joint, or gafem
• var: The variable to plot. Defaults to var = 1 the first variable of x.
• conf: Level of confidence interval to use in the Q-Q plot (0.95 by default).
• labels: Logical argument. If TRUE labels the points outside confidence interval limits.
• plot_theme: The graphical theme of the plot. Default is plot_theme = theme_metan(). For more details, see ggplot2::theme().
• band.alpha, point.alpha: The transparency of confidence band in the Q-Q plot and the points, respectively. Must be a number between 0 (opaque) and 1 (full transparency).
• fill.hist: The color to fill the histogram. Default is 'gray'.
• col.hist: The color of the border of the the histogram. Default is 'black'.
• col.point: The color of the points in the graphic. Default is 'black'.
• col.line: The color of the lines in the graphic. Default is 'red'.
• col.lab.out: The color of the labels for the 'outlying' points.
• size.lab.out: The size of the labels for the 'outlying' points.
• size.tex.lab: The size of the text in axis text and labels.
• size.shape: The size of the shape in the plots.
• bins: The number of bins to use in the histogram. Default is 30.
• which: Which graphics should be plotted. Default is which = c(1:4) that means that the first four graphics will be plotted.
• ncol, nrow: The number of columns and rows of the plot pannel. Defaults to NULL
• ...: Additional arguments passed on to the function patchwork::wrap_plots().

Examples

library(metan)
model <- performs_ammi(data_ge, ENV, GEN, REP, GY)

# Default plot
plot(model)

# Normal Q-Q plot
# Label possible outliers
plot(model,
     which = 2,
     labels = TRUE)

# Residual vs fitted,
# Normal Q-Q plot
# Histogram of raw residuals
# All in one row
plot(model,
     which = c(1, 2, 5),
     nrow = 1)

Author(s)

Tiago Olivoto tiagoolivoto@gmail.com