Struct DiffusionMaps 

pub struct DiffusionMaps { /* private fields */ }

Diffusion Maps for nonlinear dimensionality reduction

Example

use scirs2_transform::reduction::diffusion_maps::DiffusionMaps;
use scirs2_core::ndarray::Array2;

let data = Array2::<f64>::zeros((50, 10));
let mut dm = DiffusionMaps::new(2);
let embedding = dm.fit_transform(&data).expect("should succeed");
assert_eq!(embedding.shape(), &[50, 2]);

Implementations

impl DiffusionMaps

pub fn new(n_components: usize) -> Self

Create a new DiffusionMaps instance

Arguments

• n_components - Number of dimensions in the embedding

pub fn with_epsilon(self, epsilon: f64) -> Self

Set the kernel bandwidth (epsilon)

pub fn with_alpha(self, alpha: f64) -> Self

Set the anisotropy parameter (alpha)

• alpha = 0: graph Laplacian normalization
• alpha = 0.5: Fokker-Planck diffusion
• alpha = 1.0: Laplace-Beltrami (geometry-only, density-independent)

pub fn with_diffusion_time(self, t: f64) -> Self

Set the diffusion time (t)

Larger t emphasizes larger-scale structure; smaller t captures local detail.

pub fn embedding(&self) -> Option<&Array2<f64>>

Get the embedding

pub fn eigenvalues(&self) -> Option<&Array1<f64>>

Get the eigenvalues of the diffusion operator

pub fn spectral_gaps(&self) -> Option<&Array1<f64>>

Get the spectral gaps (ratio of consecutive eigenvalues)

pub fn epsilon_used(&self) -> Option<f64>

Get the epsilon (bandwidth) used

pub fn suggest_dimensionality(&self, max_dim: usize) -> Result<usize>

Automatic dimensionality selection via spectral gap

Finds the largest spectral gap (ratio lambda_{i} / lambda_{i+1}) which indicates the intrinsic dimensionality of the data.

Arguments

• max_dim - Maximum dimensionality to consider

Returns

• Suggested number of dimensions

pub fn fit<S>(&mut self, x: &ArrayBase<S, Ix2>) -> Result<()>

where S: Data, S::Elem: Float + NumCast,

Fit the Diffusion Maps model

Arguments

• x - Input data, shape (n_samples, n_features)

pub fn transform<S>(&self, x: &ArrayBase<S, Ix2>) -> Result<Array2<f64>>

where S: Data, S::Elem: Float + NumCast,

Transform data using the fitted model

pub fn fit_transform<S>(&mut self, x: &ArrayBase<S, Ix2>) -> Result<Array2<f64>>

where S: Data, S::Elem: Float + NumCast,

Fit and transform in one step

Trait Implementations

impl Clone for DiffusionMaps

fn clone(&self) -> DiffusionMaps

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for DiffusionMaps

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

Formats the value using the given formatter.