BeliefNetwork

class matchup.models.algorithms.belief_network.BeliefNetwork

Bases: matchup.models.model.IterModel

Methods Summary

generate_scores(doc_repr, float], …)	Calculate similarity between one document and query based in Belief Network IR model.
run(query, vocabulary)	Belief Network algorithm.

Methods Documentation

classmethod generate_scores(doc_repr: DefaultDict[str, float], query_repr: DefaultDict[str, float], prob_k: float) → float

Calculate similarity between one document and query based in Belief Network IR model.

Parameters:

• doc_repr – Document vector space representation
• query_repr – Query vector space representation
• prob_k – Probability that a random key ‘k’ is relevant

:return Relevant score of document based in query.

run(query: List[matchup.presentation.text.Term], vocabulary: matchup.structure.vocabulary.Vocabulary) → List[matchup.structure.solution.Result]

Belief Network algorithm.

Parameters:

• query – list of all query terms
• vocabulary – data structure that represents the vocabulary

Returns:

list of solution -> (document, score)