Struct VariableElimination 

pub struct VariableElimination {
    pub elimination_order: Option<Vec<String>>,
}

Variable elimination algorithm for exact inference.

Computes marginal probabilities by eliminating variables in a specific order.

Fields

elimination_order: Option<Vec<String>>

Elimination order (if None, uses min-degree heuristic)

Implementations

impl VariableElimination

pub fn new() -> Self

Create a new variable elimination algorithm.

pub fn with_order(order: Vec<String>) -> Self

Create with a specific elimination order.

pub fn marginalize( &self, graph: &FactorGraph, query_var: &str, ) -> Result<ArrayD<f64>>

Compute marginal for a single query variable.

pub fn marginalize_all( &self, graph: &FactorGraph, ) -> Result<HashMap<String, ArrayD<f64>>>

Compute marginals for all variables.

pub fn joint_probability( &self, graph: &FactorGraph, assignment: &HashMap<String, usize>, ) -> Result<f64>

Compute joint probability for a specific assignment.

pub fn map(&self, graph: &FactorGraph) -> Result<HashMap<String, usize>>

Compute MAP (Maximum A Posteriori) assignment using variable elimination.

Trait Implementations

impl Default for VariableElimination

fn default() -> Self

Returns the “default value” for a type.