using Word2Vec

?word2vec

word2vec("Downloads/text8", "text8-vec.txt", verbose=true)

word2phrase("Downloads/text8", "text8phrase")
word2vec("text8phrase", "text8phrase-vec.txt", verbose=true)

word2clusters("text8", "text8-class.txt", 100)

;ls

model = wordvectors("text8-vec.txt")

size(model)

words = vocabulary(model)

idx = index(model, "book")

words[idx]

get_vector(model, "one")

similarity(model, "one", "two")

similarity(model, "one", "hello")

idxs, dists = cosine(model, "paris", 10)

using Gadfly

plot(x=words[idxs], y=dists)

?analogy

indxs, dists = analogy(model, ["king", "woman"], ["man"], 8)

plot(x=words[indxs], y=dists)

?analogy_words

analogy_words(model, ["paris", "germany"], ["france"], 10)

model2 = wordvectors("text8phrase-vec.txt")

cosine_similar_words(model2, "los_angeles", 13)

model3 = wordclusters("text8-class.txt")

clusters(model3)

get_cluster(model3, "two")

get_words(model3, 39)