ML Academy · Track 3 · Large Language Models

From embeddings to topics: finding topics by clustering

Extracting topics from an unlabelled pile of text is possible. But answering "how many topics are there" without labels is not always possible.

4 steps 225 XP A free account is needed

Sources

Rousseeuw, P. J. 1987 · Silhouettes: a graphical aid to the interpretation and validation of cluster analysis · J. Comput. Appl. Math. 20
Grootendorst, M. 2022 · BERTopic: Neural topic modeling with a class-based TF-IDF procedure · arXiv:2203.05794
Chang, J. et al. 2009 · Reading Tea Leaves: How Humans Interpret Topic Models · NeurIPS 2009
von Luxburg, U. 2010 · Clustering Stability: An Overview · Foundations and Trends in ML 2(3)

ML Academy · an interactive machine learning course that runs in your browser · All lessons