vecplot(I, graph = "VEC", leg = NULL, xval = 1, sort = FALSE, data = NULL,
digits = c(1, 1), TC = 1, intbar = NULL, lty = 1, pch = 19, col = NULL,
datacol = NULL, main = "", main2 = "", Grid = 0,
xlab = "", ylab = "", zlab = "",
levels = NULL, levels2 = NULL, showlevels = FALSE,
screen = list(z = 40, x = -60), zoom = 1, cex = 1)Importance function.
VEC -- 1-D VEC curve;
VECB -- 1-D VEC curve with box plots (only valid for SA methods: "DSA", "MSA");
VEC3 -- 2-D VEC surface;
VECC -- 2-D VEC contour;
mgraphgraph="VEC" or (graph="VEC3" or "VECC"
and length(interactions)=1, see Importance). if a vector, then several VEC curves are plotted (in this case, x-axis is scaled).increasing -- sorts the first attribute (if factor) according to the response values, increasing order;
decreasing -- similar to increasing but uses reverse order;
TRUE -- similar to increasing;
increasing2 -- sorts the second attribute (for graph="VEC3" or "VECC", if factor, according to the response values), increasing order;
decreasing2 -- similar to increasing2 but uses reverse order;
FALSE -- no sort is used;
mgraphmgraphmgraphmgraphmgraphgraph="VEC" curve, can be a vector if there are several VEC curve plotsgraph="VEC"mgraphgraph="VECC"mgraphgraph="VEC3":
FALSE or TRUE -- do not (do) show the levels in x1, x2 and z axis for factor variables;
c(FALSE,FALSE,TRUE) only shows for z-axis).
VEC3 graph:
x -- assumes list(z=0,x=-90,y=0);
X -- assumes list(x=-75);
y -- assumes list(z=0,x=-90,y=-90);
Y -- assumes list(z=10,x=-90,y=-90);
z -- assumes list(z=0,x=0,y=0);
xy -- assumes list(z=10,x=-90,y=-45);
wireframe
graph="VEC3")Importance
P. Cortez and M.J. Embrechts. Using Sensitivity Analysis and Visualization Techniques to Open Black Box Data Mining Models. In Information Sciences, Elsevier, 225:1-17, March 2013.
Importance