Fixed point clusters, clusterwise regression and discriminant
plots
Description
Fuzzy and crisp fixed point cluster analysis based on
Mahalanobis distance and linear regression fixed point
clusters. Semi-explorative, semi-model-based clustering
methods, operating on n*p data, do not need prespecification of
number of clusters, produce overlapping clusters. Symmetric and
asymmetric discriminant projections separate groups optimally,
used to visualize the separation of groupings, visual cluster
validation. Clusterwise linear regression by normal mixture
modeling. Cluster validation statistics for distance based
clustering including corrected Rand index. Clusterwise cluster
stability assessment. DBSCAN clustering. Interface functions
for many clustering methods implemented in R. Note that the use
of the package mclust (called by function prabclust) is
protected by a special license, see
http://www.stat.washington.edu/mclust/license.txt, particularly
point 6.