Struct Hmm 

pub struct Hmm {
    pub pi: Vec<f64>,
    pub a: Vec<f64>,
    pub b: Vec<f64>,
    pub k: usize,
    pub m: usize,
}

A discrete HMM: k hidden states, m observation symbols.

Fields

pi: Vec<f64>

Initial state distribution π (length k).

a: Vec<f64>

Row-major k×k transition matrix A (a[i*k+j] = P(state j | state i)).

b: Vec<f64>

Row-major k×m emission matrix B (b[i*m+o] = P(symbol o | state i)).

k: usize

m: usize

Implementations

impl Hmm

pub fn new( pi: Vec<f64>, a: Vec<f64>, b: Vec<f64>, k: usize, m: usize, ) -> Result<Self, LearningError>

Construct from parameters, validating shapes (rows need not be exactly normalized but must be non-empty).

pub fn log_likelihood(&self, obs: &[usize]) -> Option<f64>

Log-likelihood log P(obs | model). None for an empty sequence or an out-of-range symbol.

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

Viterbi most-likely state path + its log-probability. None for an empty sequence or an out-of-range symbol.

Trait Implementations

impl Clone for Hmm

fn clone(&self) -> Hmm

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Hmm

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

Formats the value using the given formatter.