Evaluate Term Embeddings

Evaluate given term embeddings based on analogy_task and synonym_task. The list of test pairs should be given. The analogy_task is performed for each type of given analogies (initially, there are 7 proposed types of analogies, see terms_pairs_test). The accuracy is measured for the context size from 1 (the most restrictive) to the given n_max.

evaluate_term_embeddings(term_vectors, terms_pairs = NULL, n_max = 5L)

Arguments

term_vectors	A matrix of embeddings of the terms
n_max	A maximum number of neighbors included in the context (default: 5)
terms_pairs:	A list of lists of the test pairs, each list of pairs should contain two vectors of terms; if the list contains the element "synonym", the synonym_task is performed for these terms, for the rest pairs there is performed analogy_task

Value

A data frame of accuracies of the analogy task and the synonym task for each given list of pairs for the context from 1 to n_max.

See also

analogy_task, synonym_task, terms_pairs_test

Examples

term_vectors <- embed_terms(examinations, embedding_size = 5L,
  term_count_min = 1L)
#> Error in .subset2(public_bind_env, "initialize")(...): unused arguments (word_vectors_size = 5, vocabulary = list(c("fever", "man", "mother", "cough", "heart", "patient", "thyroid", "eye", "rhinitis", "woman", "father"), c(2, 2, 2, 3, 3, 3, 3, 4, 5, 6, 7), c(2, 2, 2, 3, 3, 3, 3, 4, 5, 6, 7)))
evaluate_term_embeddings(term_vectors, terms_pairs = terms_pairs_test)
#> Error in rownames(term_vectors): object 'term_vectors' not found