The Joint Bertin Classification Index

Computes the Joint Bertin Classification Criterion which uses joint independence as a reference for normalization.

JBCI(x, r = 1)

Arguments

x	The 3D table with non-negative entries.
r	The index of the variable which is tested for joint independence of the other two.

Details

The BCI of a 3D table but instead of the total independence case the joint independence case is used for normalization. With an optimal reordering we have JBCI(x) >= BCI(x).

Value

Numeric value in [0,1].

See also

BCI, CBCI, WBCI

Examples

# NOT RUN {
A <- optile( arsim(144*5*20,c(12,12),6,0.1) , iter = 1000)

p1 <- 0.1 + runif(5)
p1 <- p1/sum(p1)

A2 <- apply(A,1:2,function(z) rmultinom(1,z,p1))
A2 <- optile(A2, iter = 1000,return.type="table")

BCI(A)
BCI(A2)

DA2 <- subtable(A2,1:3)
names(DA2) <- c("X","Y","Z","Freq")

rmb(~Y+Z+X,data=DA2)

JBCI(A2,3)
# }