This library implements the following clustering algorithms:
- K-means
- K-medoids (using either 'Partition Around Medoids' or the 'Voronoi Iteration' algorithms)
- Agglomerative clustering (yielding dendrograms)
A basic example can be found in the test
subdirectory.
Multi-start routines are also available to pick the best out of n
initial
clusterings. At the time of writing, the implementation is entirely sequential.
- many low-hanging fruits for optimization
- implement parallel multi-start routine when multicore lands