fpc (version 2.1-10)

zmisclassification.matrix: Matrix of misclassification probabilities between mixture components

Description

Matrix of misclassification probabilities in a mixture distribution between two mixture components from estimated posterior probabilities regardless of component parameters, see Hennig (2010).

Usage

zmisclassification.matrix(z,pro=NULL,clustering=NULL,
                                      ipairs="all",symmetric=TRUE,
                                      stat="max")

Arguments

z

matrix of posterior probabilities for observations (rows) to belong to mixture components (columns), so entries need to sum up to 1 for each row.

pro

vector of component proportions, need to sum up to 1. Computed from z as default.

clustering

vector of integers giving the estimated mixture components for every observation. Computed from z as default.

ipairs

"all" or list of vectors of two integers. If ipairs="all", computations are carried out for all pairs of components. Otherwise, ipairs gives the pairs of components for which computations are carried out.

symmetric

logical. If TRUE, the matrix is symmetrised, see parameter stat.

stat

"max" or "mean". The statistic by which the two misclassification probabilities are aggregated if symmetric=TRUE.

Value

A matrix with the (symmetrised, if required) misclassification probabilities between each pair of mixture components. If symmetric=FALSE, matrix entry [i,j] is the estimated probability that an observation generated by component j is classified to component i by maximum a posteriori rule.

References

Hennig, C. (2010) Methods for merging Gaussian mixture components, Advances in Data Analysis and Classification, 4, 3-34.

See Also

confusion

Examples

# NOT RUN {
  set.seed(12345)
  m <- rpois(20,lambda=5)
  dim(m) <- c(5,4)
  m <- m/apply(m,1,sum)
  round(zmisclassification.matrix(m,symmetric=FALSE),digits=2) 
# }