shallot (version 0.4.8)

partition.confidence: Partition Confidence

Description

This function calculates the partition confidence of a partition estimate from the corresponding expected pairwise allocation matrix (EPAM).

Usage

partition.confidence(x, y)

Arguments

x

If y is not specified then x must be an object of class salso.confidence. Otherwise, x is a vector of cluster labels and y is an expected pairwise allocation matrix.

y

If y is not specified then x must be an object of class salso.confidence. Otherwise, x is a vector of cluster labels and y is an expected pairwise allocation matrix.

Value

A vector of partition confidences.

Details

The partition.confidence takes as input an object of class salso.confidence and then calculates the partition confidence from the expected pairwise allocation matrix.

The partition confidence is the average values of the EPAM for items that are clustered together. Items which are in their own subset do not contribute to partition confidence.

See Also

Other Default Mass Selection: default.mass(), mass.algorithm(), variance.ratio()

Examples

# NOT RUN {
x <- rep(c(1,2,3), times=c(2,3,5))
y <- diag(10)
y[upper.tri(y)] <- runif(45)
partition.confidence(x,y)

# }