---
-This program takes n time-series in input and is divided in two stages:
- 1) (Heuristic) k-medoids algorithm in parallel to obtain K centers, K << n
- 2) Filtering WER(?) to obtain the final k < K group representations.
+This program takes N time-series in input and is divided in two stages:
+ 1. (Heuristic) k-medoids algorithm in parallel to obtain K1 centers, K1 ≪ N
+ 2. Clustering from WER distances to obtain the final K2 < K1 group representations
----
-
-The folder "communication/" contains latex sources of the slides showed at half-project(?)
-
-The other folder contains all the R and C code, with toy datasets (real ones are not public).
+See ?epclust once the package is loaded.