I believe that the title is self-contained. Does make sense to use t-SNE and then applied HDBSCAN to cluster the data with dimensionality reduction?
Asked
Active
Viewed 243 times
1 Answers
1
tSNE is NOT a Dimensionality Reduction algorithm but a Visualization method.
The reason is that it is non-parametric and can not model a new data in the same way. What you are looking for is UMAP and yes, you can reduce dimensionality and use a clustering method to find clusters (indeed this is a common practice in clustering)
Kasra Manshaei
- 6,502
- 1
- 20
- 45