bnosac/udpipe

add dbscan together with word2vec/doc2vec/paragraph2vec

jwijffels opened this issue · 2 comments

add dbscan together with word2vec/doc2vec/paragraph2vec

doc2vec should be on cran in 2 days, so paragraph2vec (PV-DM/PV-DBOW) will be easy now
So we can now easily do https://github.com/ddangelov/Top2Vec and combine it with udpipe / sentencepiece / dbscan / uwot / word2vec / tfidf to obtain semantic topic detection. Even https://github.com/MaartenGr/BERTopic#ctfidf is in reach using golgotha