Struct linfa_hierarchical::HierarchicalCluster

pub struct HierarchicalCluster<T> { /* fields omitted */ }

Agglomerative hierarchical clustering

In this clustering algorithm each point is first considered as a seperate cluster. In 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 transformed of the similarity kernel.

Implementations

impl<F: Float> HierarchicalCluster<F>

pub fn with_method(self, method: Method) -> HierarchicalCluster<F>

Select a merging method

pub fn num_clusters(self, num_clusters: usize) -> HierarchicalCluster<F>

Stop merging when a certain number of clusters are reached

In the fitting process points are merged until a certain criterion is reached. With this option the merging process will stop, when the number of clusters drops below this value.

pub fn max_distance(self, max_distance: F) -> HierarchicalCluster<F>

Stop merging when a certain distance is reached

In the fitting process points are merged until a certain criterion is reached. With this option the merging process will stop, then the distance exceeds this value.

Trait Implementations

impl<T> Default for HierarchicalCluster<T>

Initialize hierarchical clustering, which averages in-cluster points and stops when two clusters are reached.

pub fn default() -> HierarchicalCluster<T>

Returns the "default value" for a type.

impl<'a, F: Float, T: Targets> Transformer<Dataset<Kernel<ArrayBase<ViewRepr<&'a F>, Dim<[usize; 2]>>>, T>, Dataset<Kernel<ArrayBase<ViewRepr<&'a F>, Dim<[usize; 2]>>>, Vec<usize, Global>>> for HierarchicalCluster<F>

pub fn transform(
    &self,
    dataset: Dataset<Kernel<ArrayView2<'a, F>>, T>
) -> Dataset<Kernel<ArrayView2<'a, F>>, Vec<usize>>

Perform hierarchical clustering of a similarity matrix

Returns the class id for each data point

impl<'b: 'a, 'a, F: Float> Transformer<Kernel<ArrayBase<ViewRepr<&'a F>, Dim<[usize; 2]>>>, Dataset<Kernel<ArrayBase<ViewRepr<&'a F>, Dim<[usize; 2]>>>, Vec<usize, Global>>> for HierarchicalCluster<F>

pub fn transform(
    &self,
    kernel: Kernel<ArrayView2<'a, F>>
) -> Dataset<Kernel<ArrayView2<'a, F>>, Vec<usize>>

Perform hierarchical clustering of a similarity matrix

Returns the class id for each data point