Struct DirichletProcess 

pub struct DirichletProcess {
    pub alpha: f64,
    pub assignments: Vec<usize>,
    pub cluster_counts: Vec<usize>,
    pub cluster_means: Vec<f64>,
    pub cluster_ss: Vec<f64>,
    pub n_assigned: usize,
}

Dirichlet Process mixture model using the Chinese Restaurant Process.

New data points are assigned to existing clusters with probability proportional to cluster size, or start a new cluster with probability α/(n+α).

Fields

alpha: f64

Concentration parameter α.

assignments: Vec<usize>

Cluster assignments for each observed data point.

cluster_counts: Vec<usize>

Number of points in each cluster.

cluster_means: Vec<f64>

Cluster means (updated incrementally).

cluster_ss: Vec<f64>

Cluster sum of squares (for variance estimation).

n_assigned: usize

Total number of data points assigned.

Implementations

impl DirichletProcess

pub fn new(alpha: f64) -> Self

Creates a new DirichletProcess with concentration alpha.

pub fn n_clusters(&self) -> usize

Returns the number of clusters.

pub fn crp_assign(&mut self, x: f64) -> usize

Assigns a new data point via Chinese Restaurant Process probabilities.

Returns the cluster index assigned (deterministic: picks highest prob).

pub fn stick_breaking_weights(&self, k: usize) -> Vec<f64>

Stick-breaking construction: samples first k mixture weights.

Returns weights (w_1, ..., w_k) where Σ w_i ≈ 1. Uses pseudo-random beta draws based on alpha and index.

pub fn cluster_variances(&self) -> Vec<f64>

Returns cluster variance estimates (unbiased).

pub fn expected_clusters(alpha: f64, n: usize) -> f64

Returns the expected number of clusters for n observations (approximation).

E[K_n] ≈ α ln(1 + n/α)

Trait Implementations

impl Clone for DirichletProcess

fn clone(&self) -> DirichletProcess

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for DirichletProcess

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

Formats the value using the given formatter.