Incremental nested partition method
Jyrko Correa-Morris, José Ruíz-Shulcloper, Dustin L. Espinosa-Isidrón, Aurora Pons-Porrata · Proceedings - International Conference on Pattern Recognition/Proceedings/International Conference on Pattern Recognition · 2008
In this paper, we introduce an incremental nested partition algorithm for finding the inner structuralization of dynamic datasets. Here we use three partition criteria that allow to obtain a hierarchy of clusterings. The algorithm is based on some mathematical properties, which are introduced in the paper. The experimental results over the AFP and TDT2 news collections show the usefulness of our method to reveal different levels of the information hidden in the datasets.