inferlo.NormalFactorGraphModel

class inferlo.NormalFactorGraphModel

Normal Factor Graph model.

Also known as Forney-style or Edge-Variable graphical model.

Variables correspond to edges of a graph, factors correspond to vertices of a graph.

Every variable appears in exactly two factors.

__init__(num_variables: int, domain: Domain)

Initializes Normal Factor Graph model.

Parameters:

• num_variables – Number of variables in the model.

• domain – Default domain of each variable.

Methods

__init__(num_variables, domain)	Initializes Normal Factor Graph model.
add_factor(factor)	Add a factor.
build()	Validates the model and makes it immutable.
check_built()	Checks that model is built.
draw_edge_variable_graph(ax)	Draws edge-variable graph.
draw_factor_graph(ax)	Draws the factor graph.
evaluate(x)	Returns value of non-normalized pdf in point.
from_model(original_model)	Constructs NFG model which is equivalent to given model.
get_edge_variable_graph()	Returns edge-variable graph.
get_factor_graph()	Builds factor graph for the model.
get_factors()	Returns all factors.
get_max_domain_size()	Returns the biggest domain size over all variables.
get_symbolic_variables()	Prepares variables for usage in expressions.
get_variable(idx)	Returns variable by its index.
get_variables()	Returns all variables.
infer([algorithm])	Calculates partition function.
max_likelihood(algorithm, **kwargs)	Finds the most probable state.
max_likelihood_bruteforce()	Evaluates most likely state in a very inefficient way.
part_func_bruteforce()	Evaluates partition function in very inefficient way.
sample(num_samples[, algorithm])	Generates samples.

add_factor(factor: Factor)

Add a factor.

build()

Validates the model and makes it immutable.

check_built()

Checks that model is built.

draw_edge_variable_graph(ax)

Draws edge-variable graph.

static from_model(original_model: GraphModel)

Constructs NFG model which is equivalent to given model.

If necessary, clones variables and adds factors.

Guarantees that first N variables in new model will correspond to variables in original model, where N is number of variables in original model.

get_edge_variable_graph() → Graph

Returns edge-variable graph.

get_factors() → Iterable[Factor]

Returns all factors.

infer(algorithm='auto', **kwargs)

Calculates partition function.

Available algorithms

• auto - Automatic.

• edge_elimination - Edge elimination.

Parameters:

algorithm – Which algorithm to use. String.

Returns:

Partition function.

max_likelihood(algorithm, **kwargs) → ndarray

Finds the most probable state.