Struct UMAP 

pub struct UMAP { /* private fields */ }

UMAP (Uniform Manifold Approximation and Projection) dimensionality reduction

UMAP constructs a high dimensional graph representation of the data then optimizes a low dimensional graph to be as structurally similar as possible.

Example

use scirs2_transform::UMAP;
use scirs2_core::ndarray::Array2;

let data = Array2::<f64>::zeros((50, 10));
let mut umap = UMAP::new(15, 2, 0.1, 1.0, 200);
let embedding = umap.fit_transform(&data).expect("should succeed");
assert_eq!(embedding.shape(), &[50, 2]);

Implementations

impl UMAP

pub fn new( n_neighbors: usize, n_components: usize, min_dist: f64, learning_rate: f64, n_epochs: usize, ) -> Self

Creates a new UMAP instance

Arguments

• n_neighbors - Number of neighbors to consider for local structure (typically 5-50)
• n_components - Number of dimensions in the low dimensional space (typically 2 or 3)
• min_dist - Minimum distance between points in low dimensional space (typically 0.001-0.5)
• learning_rate - Learning rate for SGD optimization (typically 1.0)
• n_epochs - Number of epochs for optimization (typically 200-500)

pub fn with_random_state(self, seed: u64) -> Self

Sets the random state for reproducibility

pub fn with_metric(self, metric: &str) -> Self

Sets the distance metric

pub fn with_negative_sample_rate(self, rate: usize) -> Self

Sets the negative sampling rate

pub fn with_spectral_init(self, use_spectral: bool) -> Self

Enable or disable spectral initialization

pub fn with_local_connectivity(self, local_connectivity: f64) -> Self

Sets the local connectivity parameter

pub fn with_set_op_mix_ratio(self, ratio: f64) -> Self

Sets the set operation mix ratio

pub fn with_spread(self, spread: f64) -> Self

Sets the spread parameter

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

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

Fits the UMAP model to the input data

Arguments

• x - The input data, shape (n_samples, n_features)

Returns

• Result<()> - Ok if successful, Err otherwise

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

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

Transforms the input data using the fitted UMAP model

Arguments

• x - The input data, shape (n_samples, n_features)

Returns

• Result<Array2<f64>> - The transformed data, shape (n_samples, n_components)

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

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

Fits the UMAP model to the input data and returns the embedding

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

Returns the low dimensional embedding

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

Returns the fuzzy simplicial set graph

Trait Implementations

impl Clone for UMAP

fn clone(&self) -> UMAP

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for UMAP

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

Formats the value using the given formatter.