atac_mapper.topic_matching.utils.infer_topic_distribution

atac_mapper.topic_matching.utils.infer_topic_distribution(tf_vector, topic_word, n_topics, n_words, n_iterations=100, tol=0.0001)

Infer topic distribution for a single document.

This function is reimplementation of reference topic inference from lda package https://lda.readthedocs.io/en/latest/

Args:

tf_vector: Term frequency vector for a single document topic_word: Topic-word matrix from reference data n_topics: Number of topics n_words: Number of words (features) n_iterations: Maximum number of iterations tol: Convergence tolerance

Returns:

topic_dist: Inferred topic distribution for the document