The Fuzzy ARTMAP neural network is a supervised pattern recognition method
based on fuzzy adaptive resonance theory (ART). It is a promising method si
nce Fuzzy ARTMAP is able to carry out on-line learning without forgetting p
reviously learnt patterns (stable learning), it can recode previously learn
t categories (adaptive to changes in the environment) and is self-organisin
g. This paper presents the application of Fuzzy ARTMAP to odour discriminat
ion with electronic nose (EN) instruments. EN data from three different dat
asets, alcohol, coffee and cow's breath (in order of complexity) were class
ified using Fuzzy ARTMAP. The accuracy of the method was 100% with alcohol,
97% with coffee and 79%, respectively. Fuzzy ARTMAP outperforms the best a
ccuracy so far obtained using the back-propagation trained multilayer perce
ptron (MLP) (100%, 81% and 68%, respectively). The MLP bring by far the mos
t popular neural network method in both the field of EN instruments and els
ewhere. These results, in the case of alcohol and coffee, are better than t
hose obtained using self-organising maps, constructive algorithms and other
ART techniques. Furthermore, the time necessary to train Fuzzy ARTMAP was
typically one order of magnitude faster than back-propagation. The results
show that this technique is very promising for developing intelligent EN eq
uipment, in terms of its possibility for on-line learning, generalisation a
bility and ability to deal with uncertainty (in terms of measurement accura
cy, noise rejection, etc.). (C) 1999 Elsevier Science S.A. All rights reser
ved.