Point to python code for synthetic hierarchical dataset.

jlmelville · Jan 12, 2025 · 94e684b · 94e684b
1 parent 2df746d
commit 94e684b
Showing 1 changed file with 3 additions and 2 deletions.
diff --git a/vignettes/articles/umap-examples.Rmd b/vignettes/articles/umap-examples.Rmd
@@ -376,8 +376,9 @@ clusters,” and each of the meso subclusters has 5 “micro clusters.” Thus,
 are 5x5x5 = 125 micro clusters total. We colored each point based on its true
 macro clusters, shading each meso cluster.". Get all of that? There's more
 detail in Appendix B of the paper and I think I was able to turn that into some
-Python that probably generated the correct data. I think it's correct. I will
-find somewhere to upload this code eventually.
+Python that probably generated the correct data. I think it's correct. For more
+details see <https://github.com/jlmelville/drnb/blob/master/notebooks/data-pipeline/synthetic-hierarchical.ipynb>.
+I have not yet summoned the enthusiasm to come up with a pure R version.
 
 | UMAP                        | t-SNE                     |
 :----------------------------:|:--------------------------: