| | |
Summary: FLIP-ECOC: A Greedy Optimization of the
ECOC Matrix
Cemre Zor¹ ², Berrin Yanikoglu¹, Terry Windeatt², Ethem Alpaydin³
¹Sabanci University, Tuzla, Istanbul, Turkey, 34956
(cemre, berrin)@sabanciuniv.edu
²Center for Vision, Speech and Signal Processing, University of Surrey, UK, GU2 7XH
t.windeatt@surrey.ac.uk
³Bogazici University, Bebek, Istanbul, Turkey, 34342
alpaydin@boun.edu.tr
Abstract. Error Correcting Output Coding (ECOC) is a multiclass
classification technique, in which multiple base classifiers (dichotomiz-
ers) are trained using subsets of the training data, determined by a preset
code matrix. While it is one of the best solutions to multiclass problems,
ECOC is suboptimal, as the code matrix and the base classifiers are not
learned simultaneously. In this paper, we show an iterative update al-
gorithm that reduces this decoupling. We compare the algorithm with
the standard ECOC approach, using Neural Networks (NNs) as the base
classifiers, and show that it improves the accuracy for some well-known
data sets under different settings.
1 Introduction
|
