[Back]

G. Pölzlbauer, T. Lidy, A. Rauber:
"Decision Manifolds: Classification Inspired by Self-Organization";
Talk: International Workshop on Self-Organizing Maps (WSOM'07), Bielefeld, Germany; 2007-09-03 - 2007-09-06; in: "6th International Workshop on Self-Organizing Maps", H. Ritter, R. Haschke (ed.); (2007), ISBN: 978-3-00-022473-7; 8 pages.

We present a classifier algorithm that
approximates the decision surface of labeled data by a
patchwork of separating hyperplanes. The hyperplanes are
arranged in a way inspired by how Self-Organizing Maps
are trained. We take advantage of the fact that the boundaries
can often be approximated by linear ones connected
by a low-dimensional nonlinear manifold. The resulting
classifier allows for a voting scheme that averages over
neighboring hyperplanes. Our algorithm is computationally
efficient both in terms of training and classification.
Further, we present a model selection framework for estimation
of the paratmeters of the classification boundary,
and show results for artificial and real-world data sets.

Decision Manifolds, supervised learning, ensemble classification

http://publik.tuwien.ac.at/files/pub-inf_5009.pdf