benchmark kmodel for consensus clustering

courtois-neuromod / dypac

Dynamic parcel aggregation for fMRI data

0 stars 2 forks source link

benchmark kmodel for consensus clustering #5

Closed pbellec closed 5 years ago

pbellec commented 5 years ago

the algorithm seems very slow, even with multithreading.

pbellec commented 5 years ago

decided to revert back to k-means. Motivation was to be able to work with booleans, and k-means seems to work, so no real reason to keep exploring k-modes.