Expand description

Hierarchical Clustering

linfa-hierarchical provides an implementation of agglomerative hierarchical clustering. In this clustering algorithm, each point is first considered as a separate cluster. During each step, two points are merged into new clusters, until a stopping criterion is reached. The distance between the points is computed as the negative-log transform of the similarity kernel.

Documentation: latest.

The big picture

linfa-hierarchical is a crate in the linfa ecosystem, a wider effort to bootstrap a toolkit for classical Machine Learning implemented in pure Rust, akin in spirit to Python’s scikit-learn.

Current state

linfa-hierarchical implements agglomerative hierarchical clustering with support of the kodama crate.


Agglomerative hierarchical clustering


Criterion when to stop merging

Error variants from parameter construction

A method for computing the dissimilarities between clusters.

Type Definitions

Simplified Result using HierarchicalError as error type