plot.fpca: Plot the results of a functional PCA

In registr: Curve Registration for Exponential Family Functional Data

Plot the results of a functional PCA

Description

S3 plot method for class fpca. Plot FPCA results by visualizing the variation of the individual FPCs around the global mean. based on an object created with function fpca_gauss, bfpca or gfpca_twoStep.

The shares of explained variance are included in the plot titles if x contains an element evalues_sum.

Usage

## S3 method for class 'fpca'
plot(
  x,
  plot_FPCs = 1:x$npc,
  sd_factor = 2,
  response_function = NULL,
  add_symbols = TRUE,
  subtitle = TRUE,
  xlim = NULL,
  ylim = NULL,
  xlab = "t [registered]",
  ylab = "y",
  ...
)

Arguments

x	Object of class "fpca".
plot_FPCs	Optional index vector of the FPCs to be plotted. Defaults to all FPCs contained in x.
sd_factor	Numeric factor with which the standard deviations of each FPC's scores are multiplied to display its variation in the plots. Defaults to 2.
response_function	Optional response function to be applied before plotting the curves. Defaults to NULL, i.e. the identity function if x$family is one of c("gaussian","binomial") or exp() if x$family is one of c("gamma","poisson").
add_symbols	Indicator if '+' and '-' symbols should be added to the plot to highlight the direction of the displayed FPCs. Defaults to TRUE.
subtitle	If TRUE (default) the parameter sd_factor is displayed in the plot subtitle.
xlim, ylim	Optional numeric vectors with limits for the x and y axis.
xlab, ylab	Optional titles for the x and y axis.
...	Additional arguments passed to theme.

Value

@return If multiple FPCs are plotted, returns a grid of ggplot plots, created with cowplot::plot_grid. If only one FPC is plotted, returns a single ggplot plot.

Author(s)

Alexander Bauer [e-mail]

Examples

data(growth_incomplete)

fpca_obj = fpca_gauss(Y = growth_incomplete, npc = 2)
if (requireNamespace("ggplot2", quietly = TRUE) &&
requireNamespace("cowplot", quietly = TRUE)) {
library(ggplot2)
plot(fpca_obj)
}

registr documentation built on Oct. 3, 2022, 1:05 a.m.