Struct HiddenMarkovModel 

pub struct HiddenMarkovModel {
    pub n_states: usize,
    pub initial: Vec<f64>,
    pub transition: Vec<Vec<f64>>,
    pub emission_mean: Vec<f64>,
    pub emission_var: Vec<f64>,
}

A Hidden Markov Model with discrete hidden states and Gaussian emissions.

Supports:

• Forward algorithm (likelihood computation)
• Viterbi algorithm (MAP state sequence)
• Baum-Welch EM (parameter learning)

Fields

n_states: usize

Number of hidden states.

initial: Vec<f64>

Initial state distribution π.

transition: Vec<Vec<f64>>

Transition matrix A[i][j] = P(s_t=j | s_{t-1}=i).

emission_mean: Vec<f64>

Emission mean for each state.

emission_var: Vec<f64>

Emission variance for each state.

Implementations

impl HiddenMarkovModel

pub fn new( n_states: usize, initial: Vec<f64>, transition: Vec<Vec<f64>>, emission_mean: Vec<f64>, emission_var: Vec<f64>, ) -> Self

Creates a new HiddenMarkovModel with given parameters.

pub fn uniform(n_states: usize) -> Self

Creates a uniform HMM with n_states states.

pub fn forward(&self, observations: &[f64]) -> f64

Forward algorithm: returns log-likelihood of observation sequence.

pub fn viterbi(&self, observations: &[f64]) -> Vec<usize>

Viterbi algorithm: returns the most likely state sequence.

pub fn baum_welch(&mut self, observations: &[f64], n_iter: usize) -> Vec<f64>

Baum-Welch EM algorithm for parameter estimation.

Returns the log-likelihood at each iteration.

Trait Implementations

impl Clone for HiddenMarkovModel

fn clone(&self) -> HiddenMarkovModel

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for HiddenMarkovModel

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

Formats the value using the given formatter.