Open saurabh0512 opened 2 years ago
Yes, DBSCan and HDBSCAN can be used.
Also have a look at: https://www.sbert.net/examples/applications/clustering/README.html
Hi, thanks for quick reply. I have two follow-up questions. Since absolute value of cosine similarity cannot tell us anything about similarity of two sentences, how can we justify using clustering algorithms? Also how to bring embeddings of lexically diverse sentences nearer, since right now clustering is showing the sentences have token overlaps?
Hi, I have a problem statement which need similar sentences to be grouped together. Can I use clustering algorithms like DBSCAN, HDBSAN to cluster the embeddings together?