ggscreeplot: Screeplot for Principal Components

In ggbiplot: A Grammar of Graphics Implementation of Biplots

Screeplot for Principal Components

Description

Produces scree plots (Cattell, 1966) of the variance proportions explained by each dimension against dimension number from various PCA-like dimension reduction techniques.

Usage

ggscreeplot(
  pcobj,
  type = c("pev", "cev"),
  size = 4,
  shape = 19,
  color = "black",
  linetype = 1,
  linewidth = 1
)

Arguments

pcobj	an object returned by prcomp, princomp, PCA, dudi.pca, or lda
type	the type of scree plot, one of c('pev', 'cev'). 'pev' plots the proportion of explained variance, i.e. the eigenvalues divided by the trace. 'cev' plots the cumulative proportion of explained variance, i.e. the partial sum of the first k eigenvalues divided by the trace.
size	point size
shape	shape of the points. Default: 19, a filled circle.
color	color for points and line. Default: "black".
linetype	type of line
linewidth	width of line

Value

A ggplot2 object with the aesthetics x = PC, y = yvar

References

Cattell, R. B. (1966). The Scree Test For The Number Of Factors. Multivariate Behavioral Research, 1, 245–276.

Examples

data(wine)
wine.pca <- prcomp(wine, scale. = TRUE)
ggscreeplot(wine.pca)

# show horizontal lines for 80, 90% of cumulative variance
ggscreeplot(wine.pca, type = "cev") +
  geom_hline(yintercept = c(0.8, 0.9), color = "blue") 

# Make a fancy screeplot, higlighting the scree starting at component 4
data(crime)
crime.pca <- 
  crime |> 
  dplyr::select(where(is.numeric)) |>
  prcomp(scale. = TRUE)
  
(crime.eig <- crime.pca |> 
   broom::tidy(matrix = "eigenvalues"))

ggscreeplot(crime.pca) +
  stat_smooth(data = crime.eig |> dplyr::filter(PC>=4), 
              aes(x=PC, y=percent), method = "lm", 
              se = FALSE,
              fullrange = TRUE) 

ggbiplot documentation built on June 22, 2024, 11:05 a.m.