rect.box.idx {rflowcyt} | R Documentation |
The boundaries of a rectangle are determined from a vector of positional indicies 'box.idx' and the given variables, 'x1' and 'x2'. This box is then displayed on the existing plot.
rect.box.idx(x1, x2, box.idx = NULL, original.data.idx = 1:length(x1), border = "black", lwd = 3, ...)
x1 |
vector of values for variable 1 |
x2 |
vector of values for variable 2 |
box.idx |
vector of positional indicies that indicate the box to be shown |
original.data.idx |
positional values of the current 'x1' and 'x2' observations |
border |
the color of the outline of the box or rectangle |
lwd |
the width of the lines of the box |
... |
other options in rect |
This function would be coupled with the use of ContourScatterPlot to show the boxes obtained by 'do.PRIM' (Patient Rule Induction Method) from the rfcprim package. PRIM is a semi-automated bump-hunting program.
A.J. Rossini and J.Y. Wan
See details in rfcprim
ContourScatterPlot
, rfcprim library
if (require(rfcdmin)){ data(PRIM.example.data) if (require(rfcprim)){ ## only the peeling step is implemented out.peel <- peel.step(X.PRIM, Y.PRIM) if (interactive()){ ContourScatterPlot(X.PRIM[,1], X.PRIM[,2], status=Y.PRIM, main="z statistic", xlab=col.nm[4], ylab=col.nm[5], image.col=heat.colors(20),plot.legend.CSP=TRUE) ## the Green box is the initial estimate of the first rule ## after the peeling step rect.box.idx(out.peel@best.box.idx, X.PRIM[,1], X.PRIM[,2], border="green") } } }