Clustering
C++20 header-only: DBSCAN, HDBSCAN, k-means.
Loading...
Searching...
No Matches
clustering::hdbscan::MstOutput< T > Struct Template Reference

Frozen output contract of every MST backend. More...

#include <clustering/hdbscan/mst_output.h>

Collaboration diagram for clustering::hdbscan::MstOutput< T >:
[legend]

Public Attributes

std::vector< MstEdge< T > > edges
 The N - 1 MST edges, in insertion order.
NDArray< T, 1 > coreDistances {std::array<std::size_t, 1>{0}}
 Per-point core distance (length N; self-excluded kNN distance at minSamples).

Detailed Description

template<class T>
struct clustering::hdbscan::MstOutput< T >

Frozen output contract of every MST backend.

The MST boundary is the one axis of variation across backends; everything downstream (single-linkage tree, condensed tree, cluster extraction, outlier scoring) is monomorphic and reads from this shape. Fields default to well-defined empty values so a MstOutput produced by the default constructor is already in a valid "no fit yet" state.

Template Parameters
TElement type of the point cloud.

Definition at line 41 of file mst_output.h.

Member Data Documentation

◆ coreDistances

template<class T>
NDArray<T, 1> clustering::hdbscan::MstOutput< T >::coreDistances {std::array<std::size_t, 1>{0}}

Per-point core distance (length N; self-excluded kNN distance at minSamples).

Definition at line 45 of file mst_output.h.

◆ edges

template<class T>
std::vector<MstEdge<T> > clustering::hdbscan::MstOutput< T >::edges

The N - 1 MST edges, in insertion order.

Definition at line 43 of file mst_output.h.


The documentation for this struct was generated from the following file: