what are some function/package in R/python to find similarity of individual words not in the context of sentences? As a novice, I searched and found methods like adist, cosine similarity, word2vec, but not sure what are the supposed use case for each. Appreciate any advice.
1 Answer
If your intent is to find compare similarity in meaning, word2vec is the only appropriate choice. adist measures the edit distance between two words, and cosine similarity compares the similarity of two documents (treating the documents as bags of words). On the other hand, transforming a word into its word2vec embedding captures some of the latent meaning of the word. See this for a detailed explanation of how this is done.