Struct TSNE

pub struct TSNE { /* private fields */ }

t-SNE (t-distributed Stochastic Neighbor Embedding) for dimensionality reduction

t-SNE is a nonlinear dimensionality reduction technique well-suited for embedding high-dimensional data for visualization in a low-dimensional space (typically 2D or 3D). It models each high-dimensional object by a two- or three-dimensional point in such a way that similar objects are modeled by nearby points and dissimilar objects are modeled by distant points with high probability.

Implementations

impl TSNE

pub fn new() -> Self

Creates a new t-SNE instance with default parameters

pub fn with_n_components(self, n_components: usize) -> Self

Sets the number of components in the embedded space

pub fn with_perplexity(self, perplexity: f64) -> Self

Sets the perplexity parameter

pub fn with_early_exaggeration(self, early_exaggeration: f64) -> Self

Sets the early exaggeration factor

pub fn with_learning_rate(self, learning_rate: f64) -> Self

Sets the learning rate for gradient descent

pub fn with_max_iter(self, max_iter: usize) -> Self

Sets the maximum number of iterations

pub fn with_n_iter_without_progress( self, n_iter_without_progress: usize, ) -> Self

Sets the number of iterations without progress before early stopping

pub fn with_min_grad_norm(self, min_grad_norm: f64) -> Self

Sets the minimum gradient norm for convergence

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

Sets the metric for pairwise distance computation

pub fn with_method(self, method: &str) -> Self

Sets the method for dimensionality reduction

pub fn with_init(self, init: &str) -> Self

Sets the initialization method

pub fn with_angle(self, angle: f64) -> Self

Sets the angle for Barnes-Hut approximation

pub fn with_n_jobs(self, n_jobs: i32) -> Self

Sets the number of parallel jobs to run

• n_jobs = -1: Use all available cores
• n_jobs = 1: Use single-core (disable multicore)
• n_jobs > 1: Use specific number of cores

pub fn with_verbose(self, verbose: bool) -> Self

Sets the verbosity level

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

Sets the random state for reproducibility

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

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

Fit t-SNE to input data and transform it to the embedded space

Arguments

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

Returns

• Result<Array2<f64>> - Embedding of the training data, shape (n_samples, n_components)

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

Returns the embedding after fitting

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

Returns the KL divergence after optimization

pub fn n_iter(&self) -> Option<usize>

Returns the number of iterations run

Trait Implementations

impl Default for TSNE

fn default() -> Self

Returns the “default value” for a type.