Struct BayesianNetwork 

pub struct BayesianNetwork {
    pub dag: DAG,
    pub cpds: Vec<Box<dyn CPD>>,
}

A Bayesian Network: a DAG with a CPD for each node.

Fields

dag: DAG

The underlying DAG.

cpds: Vec<Box<dyn CPD>>

One CPD per node (indexed by node index).

Implementations

impl BayesianNetwork

pub fn new(dag: DAG, cpds: Vec<Box<dyn CPD>>) -> Result<Self, StatsError>

Create a new BayesianNetwork from a DAG and CPDs.

The cpds slice must have exactly dag.n_nodes elements, where cpds[i] is the CPD for node i.

pub fn joint_probability(&self, assignment: &[usize]) -> Result<f64, StatsError>

Compute the joint probability P(X = assignment) for a complete assignment.

P(X) = ∏_i P(X_i | pa(X_i))

pub fn cardinality(&self, node: usize) -> usize

Node cardinality.