How to cluster spacy vectors (word embedding) into groups using Annoy or other similar algorithms

Asked Jun 27 '20 at 08:46

Active Jun 27 '20 at 09:08

Viewed 1,246 times

I have a list of words whose vector embeddings I got by using spacy's pre trained model en_core_web_lg.

My questions are two fold

Can these word vectors be fed into Annoy like algorithm? Can I get say 20 groups, each group containing around 100 words?

I would be grateful for any resources/code.

edited Jun 27 '20 at 09:08

asked Jun 27 '20 at 08:46

user10083444

What have you tried so far? – chefhose Jun 27 '20 at 09:51
@chefhose Well I have tried this illustration [https://radimrehurek.com/gensim/auto_examples/tutorials/run_annoy.html#sphx-glr-auto-examples-tutorials-run-annoy-py] but the point is I know how to get similar words. I want to know how to cluster similar words into 1 group. – user10083444 Jun 27 '20 at 11:42

0 Answers0