msr (version 0.4.4)

plot.msc: Visualization of the Morse-Smale Summary of High-Dimensional Scalar Functions

Description

Visualize the Morse-Smale summary description of a high-dimensional scalar function $ y = f(x)$ with parameters $x in R^n$. For each partition of the Morse-Smale complex, an inverse regression curve is computed that summarizes the domain in that partition. This forms a network of regression curves that connect the extremal points of the function. This network is then embedded in 2D for visualization. The function value of the regression curves is encoded by color and by height in the 3rd dimension for each regression curve. Optional tubes around the regression curves indicate the standard deviation along the curve, representing the approximate extent of the partition. An additional window plots the regression curve for each parameter in $x$, which allows to examine the behaviour of each partition. Users can select, by mouse-click on the corresponding regression curve, which partitions the plots of the underlying parameters will be shown. In addition, a subset of the parameters $x$ can be selected using mscPlot$plotList.

Usage

"plot"(x, drawStdDev=FALSE, span=0.5, nsamples=50,
plot=TRUE, colorMap=0, ...)
"plot"(x, drawStdDev=FALSE, span=0.5, nsamples=50,  plot=TRUE, colorMap=0, ...)
"plot"(x, drawStdDev=FALSE, span=0.5, nsamples=50,
plot=TRUE,  colorMap=0, ...)
"plot"(x, drawStdDev=FALSE, axesOn=TRUE, ...)

Arguments

x

The Morse-Smale complex object for plot.msc or the mscPlot object for plot.mscPlot.

drawStdDev

Draw the standard deviation tubes around the plots (default FALSE).

axesOn

Draw the alignment axes (default TRUE).

nsamples

NUmber of samples for piecewise linear approximation to regression curve in each partition

span

Span argument of loess for computIng regression curves

plot

Show visualization (TRUE) or just return the plotting object (FALSE)

colorMap

The choice of colormap. 0 = Blue-Green-Red, 1 = Blue-White-Red, 2 = Purple-White-Green

...

additional args have no effect

Value

An object of class mscPlot is used to plot the Morse-Smale summary and allows to manipulate the plotting behaviour. The object mscPlot has the following components:

geom

which describes the geometry of the summary.

scene

which describes the components of the visualization scene.

plotList

which defines the specific plots to compare. This is a list containing the plot numbers.

References

[1] Samuel Gerber and Kristin Potter The Morse-Smale Complex for Data Analysis, Journal of Statistical Software, 2012, vol. 50, no. 2, pp 1-22 [2] Samuel Gerber, Oliver Ruebel Peer-Timo Bremer, Valerio Pascucci, Ross Whitaker, Morse-Smale Regression, Journal of Computational and Graphical Statistics, 2012

[3] Samuel Gerber, Peer-Timo Bremer, Valerio Pascucci, Ross Whitaker, Visual Exploration of High Dimensional Scalar Functions, IEEE Transactions on Visualization and Computer Graphics, vol. 16, no. 6, pp 1271-1280, Nov.-Dec. 2010.

See Also

msc.nn, msc.nn.svm, msc.nn.kd

Examples

data(fourpeaks)
d <- fourpeaks()

#Build Morse-Smale complex of d
ms <- msc.nn(y=d[,1], x=d[, 2:3], pLevel=0.1, knn = 15)

# Plot the ms summary
p <- plot(ms)

# Change the plots to compare
p$plotList <- c(1)

# Re-plot the summary
plot(p)