While the three consensus functions proposed are designed to also work with soft clustering inputs, we have not conducted an experimental study to evaluate their behavior in that scenario. Possibly, soft clustering ensembles will perform superior to hard clustering ensembles since the confidences on cluster memberships can be leveraged to combine clusterings in a risk-aware fashion.
One can also extend the combiners to output soft consensus clusterings. In fact, the Meta-CLustering Algorithm (MCLA) already computes soft consensus clusterings, and the Cluster-based Similarity Partitioning Algorithm (CSPA) and the HyperGraph Partitioning Algorithm (HGPA) can be extended to do so.