Fig 1.
A basic federated clustering framework.
Fig 2.
The framework of the proposed FKmeansCB.
Table 1.
Dataset features.
Table 2.
Comparison between the proposed algorithm and other clustering methods on six datasets (parameter values are in the parentheses).
Fig 3.
Running time comparison.
Fig 4.
The influence of parametersk1 and k2 on clustering results.
Fig 5.
Clustering results under different numbers of clients.
Table 3.
The influence of parameters privacy budget () on clustering results (ARI).
Fig 6.
Clustering results on R15 and MNIST by using FKmeansCB.