Clustering on 20 Newsgroups (subsampled)
1Normalized Clustering LossBanditPAM++
Evaluation Results
| Method | Links | |
|---|---|---|
| BanditPAM++Dataset Size (n)=10,000, Distance metric=cosine, k=52023.10 | 1 | |
| BanditPAM++Dataset Size (n)=15,000, Distance metric=cosine, k=52023.10 | 1 | |
| BanditPAM++Dataset Size (n)=20,000, Distance metric=cosine, k=52023.10 | 1 | |
| BanditPAM++Dataset Size (n)=25,000, Distance metric=cosine, k=52023.10 | 1 | |
| BanditPAM++Dataset Size (n)=30,000, Distance metric=cosine, k=52023.10 | 1 |