pairwise association matrix

Computes pairwise BCI values via qBCI.

cmat(x, sort = TRUE, crit = BCI, k = 5, iter = 20,
p = NULL,  jitter = TRUE, freqvar = NULL, diag = NULL,
fun = "BCC", foreign = NULL)

Arguments

x	A data.frame with factor variables or numeric variables which will be transformed to ordinal interval variables via cut. The breakpoints are quantiles of the variables such that for each pair of numeric variables the expected number of observations in each combination of intervals is at least k.
sort	Whether or not to sort the pairwise tables via optile.
crit	The criterion function, e.g. kendalls, BCI, WBCI or wdcor.
k	The minimum expected number for each cell after quantile binning. See also qBCI.
iter	An optile parameter.
p	The quantile distance. See qBCI.
jitter	Whether or not to use jittering in order to avoid ties. This is equivalent to a random assignment of ranks to observations with the same value.
freqvar	Optional weights, e.g. a frequency variable.
diag	An optional value for the diagonal. Avoids unnecessary function calls for the diagonal elements. E.g. diag = 0 for crit = BCI or diag = 1 for crit = kendalls makes sense.
fun	See optile.
foreign	See optile.

Details

Uses pairwise complete cases only!

Value

A symmetric matrix.

See also

qBCI, See wdcor.

Examples

# NOT RUN {
m1 <- cmat(olives)
fluctile(1 - m1,shape="o")
# }