Struct KernelKmeansResult 

#[non_exhaustive]
pub struct KernelKmeansResult {
    pub cluster: Vec<usize>,
    pub inertia: f64,
    pub iter: usize,
    pub converged: bool,
    pub n_init_best: usize,
    /* private fields */
}

Result of kernel_kmeans_fd.

Carries the cluster assignments and fit diagnostics plus the internal state KernelKmeansResult::predict needs (the fitted GakGramTrain, the per-cluster within-cluster kernel sums, and the cluster sizes). There is no centroid / centers field: kernel k-means has no explicit centroid curve.

Fields (Non-exhaustive)

Non-exhaustive structs could have additional fields added in future. Therefore, non-exhaustive structs cannot be constructed in external crates using the traditional Struct { .. } syntax; cannot be matched against without a wildcard ..; and struct update syntax will not work.

cluster: Vec<usize>

Cluster assignment for each training curve (values in 0..n_clusters).

inertia: f64

Total feature-space inertia Σ_i d²(i, cluster[i]) of the returned run.

iter: usize

Number of Lloyd iterations the winning restart ran.

converged: bool

Whether the winning restart converged before hitting max_iter.

n_init_best: usize

Index of the restart (0-based) that produced this result.

Implementations

impl KernelKmeansResult

pub fn n_clusters(&self) -> usize

The number of clusters (k).

pub fn predict(&self, new_data: &FdMatrix) -> Result<Vec<usize>, FdarError>

Assign new (out-of-sample) curves to the fitted clusters.

Builds the normalized cross-Gram Kcross = gak_gram_predict(train, new_data) (n_test × n_train) and assigns each new curve t to the cluster minimizing

d²(t, k) = 1 − (2/|C_k|)·Σ_{j∈C_k} Kcross[t,j] + within_k

where 1 = k(test,test) (normalized GAK), and within_k, |C_k| are the fitted training quantities. The same σ and normalization as the fit are reused (via gak_gram_predict) — nothing is re-estimated on the test set.

Errors

Propagates gak_gram_predict errors: FdarError::InvalidDimension if new_data is empty or its evaluation-grid width differs from the training set’s, or FdarError::InvalidParameter if the stored σ is invalid.

Trait Implementations

impl Clone for KernelKmeansResult

fn clone(&self) -> KernelKmeansResult

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for KernelKmeansResult

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl PartialEq for KernelKmeansResult

fn eq(&self, other: &KernelKmeansResult) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for KernelKmeansResult