Dypac
Dypac uses a simple two level clustering, one on sliding time windows, and one on indicator functions of parcels aggregated over many windows. For the sliding windows, k means runs directly on the time series generating a family of parcels, which are represented with one-hot encoders. The second-level aggregation clustering procedure gives the stable maps.
Désirée Lussier-Lévesque
Data Scientist / Artificial Intelligence Developer
I am currently a Data Scientist and Artificial Intelligence Developer with 11+ years research experience and 6+ years exerpience in AI.