Modeling social media topics on Twitter with sentence embeddings
Transform each tweet into a sentence embedding and group vectors by clusters.
Find the most relevant words in each clusters.
Create topic embedding based on the weights assigned by the most relevant words in the cluster
Some examples:
