[+] Scree plot with indicated information dimension (ggplot2)

Plot a screeplot and with indicated information dimension.

qplot_infoDim(
  obj,
  n.comp.SHOW = 20,
  selected = NA,
  Title = "Scree Plot",
  y.log = TRUE,
  show.legend = TRUE,
  legend.position = c(0.8, 0.8),
  ggtheme = theme_bw()
)

qplot_screeplot(obj, ...)

qplot_scree(obj, ...)

Arguments

obj

Either a matrix (rows = observations, columns = variables; to be passed to infoDim) or an obj (a list) generated by function infoDim.

n.comp.SHOW

A number of components to show on x axis, default is 20. This number is corrected if (a) vector of eigenvalues is smaller than 20 or (b) information dimension is higher than 15.

selected

A number of components sellected, will be plotted as a separate vertical line (optional parameter).

Title

The main title of the plot.

y.log

Logical. If TRUE (default) the scale of y axis is logarythmic.

show.legend

Logical. If TRUE (default) the legend is displayed.

legend.position

the position of the legend ("none", "left", "right", "bottom", "top", or two-element numeric vector).

ggtheme

A function of ggplot2 theme to apply (e.g.: ggtheme). Default is theme_bw().

...

(The same parameters as above).

Value

A scree plot : plot which helps to determine the number of nenessary components (e.g. for PCA).

(A "ggplot" object.)

Note

http://www.originlab.com/doc

See also

Other spHelper plots: check_palette(), layer_spRangeMean(), plot_colors(), plot_hyPalette(), plot_spCompare(), plot_spDiff(), plot_spDistribution(), qplot_confusion(), qplot_crosstab(), qplot_kAmp(), qplot_kSp(), qplot_prediction(), qplot_spRangeCenter(), qplot_spRangeMedian(), qplot_spStat(), qplot_spc(), rmExpr(), rm_stripes(), stat_chull()

Other component analysis / factorisation related functions in spHelper: getScores(), infoDim(), plot_spDiff(), qplot_kAmp(), qplot_kSp(), qplot_spc(), reconstructSp(), sortLoadings(), unipeak(), whichOutlier()

Other information dimension functions: infoDim()

Examples

 data(Spectra2, package ="spHelper")
 qplot_infoDim(Spectra2)


# ------------------------------------------------------
 my_matrix <- matrix(rexp(200, rate=.1), ncol=20)
 my_result <- infoDim(my_matrix)

 # Investigate the result
 str(my_result)
#> List of 5
#>  $ dim        : num 8
#>  $ exactDim   : num 7.91
#>  $ explained  : num [1:10] 0.307 0.1588 0.1051 0.0882 0.0814 ...
#>  $ eigenvalues: num [1:10] 162.7 84.2 55.7 46.7 43.1 ...
#>  $ n.comp     : int [1:10] 1 2 3 4 5 6 7 8 9 10
#>  - attr(*, "class")= chr [1:2] "list" "infoDim"
 my_result$exactDim
#> [1] 7.907656
 my_result$dim
#> [1] 8

 #Plot
 my_plot <- qplot_infoDim(my_result)
 my_plot



qplot_infoDim(my_matrix)