plot.varySmooth: Plot Results from GSLCCA Run with Different Smoothing...
In gslcca: Generalized Semi-Linear Canonical Correlation Analysis

Description Usage Arguments Details Author(s) See Also Examples

Plot the optimised value, ‘signatures’ (loadings of the Y matrix) or (projected) observed and fitted values (Y and X scores) for a varySmooth object.

## S3 method for class 'varySmooth'
plot(x, type = "opt", series = x[[1]]$treatment,
  subject = levels(x[[1]]$subject), ask = dev.interactive(),
  main = NULL, xlab = NULL, ylab = NULL, col = NULL, lty = 1, lwd = 1.5, 
  pch = NULL, space = "bottom", corner = NULL, columns = 2, ...)

`x`	an object of class `"varySmooth"` as returned by `varySmooth`
`type`	the type of plot: either `"opt"`, `"signature"`, `"fitted"`, or `"scores"`.
`series`	if `x[[1]]$treatment` is `NULL` and there are multiple time series for each subject, a factor to differentiate these series.
`subject`	the subjects to include in the plot, specified by levels of `x$subject`.
`ask`	logical: if `TRUE`, the user is asked before each plot, see `par(ask=.)`. Ignored when only one plot is produced.
`main`	an overall title for the plot.
`xlab`	a title for the x axis.
`ylab`	a title for the y axis.
`col`	a vector specifying a colour palette to use for the lines/points. By default the palette is specified by `rainbow_hcl` if available, otherwise the standard plotting colours are used.
`lty`	a vector of line types, see `par`.
`lwd`	a vector of line widths, see `par`.
`pch`	a vector of plotting characters, see `points`.
`space`	when `type = "fitted"`, the space outside the plot region in which to put the legend, see `xyplot`. Overrides the `corner` argument.
`corner`	when `type = "fitted"`, the corner inside the plot region in which to put the legend, see `xyplot`.
`columns`	when `type = "fitted"`, how many columns to use to format the legend, see `xyplot`.
`...`	further arguments passed to `xyplot`.

For type = "opt", the value of the optimisation criterion, log(1 - R^2) is plotted against the number of roots used in the GSLCCA, trellised by subject.

For type = "signature", the loadings of the Y matrix from the GSLCCA are displayed for each subject, trellised by the number of roots used in the GSLCCA. The arguments col and lty control the line style and legend text for each subject.

For type = "fitted", the X scores are displayed, trellised by the number of roots used in the GSLCCA. If there are multiple X scores for each time point, then a line is displayed for each times series specified by series (usually the treatment factor used in GSLCCA) and the arguments col, lty and pch control the display for each series. By default, grey or black is used to display values corresponding to the reference treatment (specified by x$ref).

For type = "scores" the X scores are displayed as for type = "fitted" and points are added for the Y scores, again coloured by series.

Heather Turner

gslcca, varySmooth

data(clonidine)

### Smoothed data - automatically select number of roots
result <- gslcca(spectra, "Critical Exponential",
    time = Time, treatment = Treatment, subject.smooth = TRUE,
    data = clonidine, subset = Rat == "42")

### Vary number of roots 
multiRoots <- varySmooth(result, 2:15)

## plot optimised value
plot(multiRoots, "opt")

## plot fitted values
plot(multiRoots, "fitted")

## plot signature
plot(multiRoots, "signature")