> library(text2vec)
> library(tm)
> product.dat <- read.csv("C:/DAR 2017/product_descriptions.csv",stringsAsFactors=FALSE)
> txt <- product.dat[,2]
> txt <- removePunctuation(txt)
> txt <- tolower(txt)
> it = itoken(txt, tolower, word_tokenizer, n_chunks = 10)
> vocab = create_vocabulary(it)

 
> dim(vocab)
[1] 364758      3

> vocab = prune_vocabulary(vocab, term_count_min = 10, doc_proportion_max = 0.8,doc_proportion_min = 0.001, vocab_term_max = 20000)
> dim(vocab)
[1] 6563    3

> vocab[1:3,]
Number of docs: 124428 
0 stopwords:  ... 
ngram_min = 1; ngram_max = 1 
Vocabulary: 
   term term_count doc_count
1: with     250258     98674
2:   of     229820     91535
3:   is     192208     87733


> vectorizer = vocab_vectorizer(vocab)
> tcm = create_tcm(it, vectorizer, skip_grams_window = 5L)
> glove = GlobalVectors$new(rank = 50, x_max = 10)
> wv_main = glove$fit_transform(tcm, n_iter = 50, convergence_tol = 0.01)
> wv_context = glove$components
> word_vectors = wv_main + t(wv_context)
> word_vectors[1:3,1:5]
           [,1]       [,2]        [,3]       [,4]        [,5]
with  0.1572956 -0.4986556 -0.30388024 -0.5208205 -0.49814761
of    0.5462567 -0.3089656 -0.40487972  0.3400864 -0.16796595
is   -0.3926085 -0.9640168  0.08781935  0.3989814 -0.08923554

> screw = word_vectors["screw", , drop = FALSE]
> cos_sim = sim2(x = word_vectors, y = screw, method = "cosine", norm = "l2")
> head(sort(cos_sim[,1], decreasing = TRUE), 5)
    screw    screws      head       hex     bolts 
1.0000000 0.7952427 0.6618135 0.6532096 0.6130304 

> design = word_vectors["design", , drop = FALSE]
> cos_sim = sim2(x = word_vectors, y = design, method = "cosine", norm = "l2")
> head(sort(cos_sim[,1], decreasing = TRUE), 5)
   design        an    unique       its     style 
1.0000000 0.7495426 0.7422682 0.7408201 0.7407434 

> cos_sim = sim2(x = word_vectors, method = "cosine", norm = "l2")
> dim(cos_sim)
[1] 6563 6563

> cos_sim[1:5,1:5]
          with        of        is        or      your
with 1.0000000 0.7087212 0.7530961 0.5669205 0.5603469
of   0.7087212 1.0000000 0.6033791 0.4730575 0.5919557
is   0.7530961 0.6033791 1.0000000 0.5426918 0.4613141
or   0.5669205 0.4730575 0.5426918 1.0000000 0.6112975
your 0.5603469 0.5919557 0.4613141 0.6112975 1.0000000

> cos_sim["of","is"]
[1] 0.6033791

> max(cos_sim["of",c("is","with")])
[1] 0.7087212

# maar let op!
> cos_sim["hahaha","is"]
Error in cos_sim["hahaha", "is"] : subscript out of bounds