zifn / CS289_Coltrims_ML

repo for the final project of berkeley CS289
0 stars 1 forks source link

Soft Clustering #30

Open sajantanand opened 3 years ago

sajantanand commented 3 years ago

Do soft clustering to get probability of class assignments from clustering rather than definitive assignments. This will allow us to calculate a more complete cross-entropy. Helpful links from googling are below.

https://hdbscan.readthedocs.io/en/latest/soft_clustering_explanation.html https://github.com/sK4rdell/soft-kmeans/blob/master/kmeans.py