Struct MarkovChain 

pub struct MarkovChain {
    pub p: Vec<f64>,
    pub n_states: usize,
}

Discrete-time finite Markov chain with transition matrix P.

Fields

p: Vec<f64>

Transition matrix P (n x n), row-stochastic.

n_states: usize

Number of states.

Implementations

impl MarkovChain

pub fn new(p: Vec<f64>, n_states: usize) -> Self

Create a Markov chain from a row-stochastic transition matrix (n x n).

pub fn simulate(&self, init: usize, n_steps: usize, seed: u64) -> Vec<usize>

Simulate a trajectory of length n_steps starting from state init.

pub fn stationary_distribution(&self, tol: f64, max_iter: usize) -> Vec<f64>

Compute the stationary distribution by iterating the distribution vector.

Returns the vector pi such that pi P = pi.

pub fn n_step_matrix(&self, n: usize) -> Vec<f64>

Compute the n-step transition matrix P^n.

pub fn mixing_time(&self, init: usize, eps: f64) -> usize

Estimate the mixing time: smallest t such that ||(pi_0 P^t) - pi||_1 < eps.

pub fn empirical_transition_matrix( trajectory: &[usize], n_states: usize, ) -> Vec<f64>

Compute the empirical transition counts from a trajectory.