package prbnmcn-clustering

You can search for identifiers within the package.

in-package search v0.2.0

prbnmcn-clustering
- Clustering
  - Agglomerative
    
    Element_set
    
    Make
    
    E
    
    S
  - Helpers
  - Intf
    
    Metric
  - K_means
    
    Element
    
    Make
    
    E
  - K_medoids
    
    Make
    
    E
  - MultiStart

Legend:
Library
Module
Module type
Parameter
Class
Class type

type t

t is the type of (multi)-sets of elements.

type elt = E.t

elt is the type of elements to be clustered.

val singleton : elt -> t

singleton x is the cluster containing x as only element.

val dist : t -> t -> float

The user should provide dist, aa distance function on clusters. A typical (costly) choice is the Hausdorff distance (see e.g. the gromov package). Other, non-metric choices are sometimes used.

val join : t -> t -> t

One should be able to "join" clusters. This can be multiset union, set union or any sensible overapproximation - the algorithm will work anyway (think for instance of convex hulls in R^n)