Toward Sensor to Reproduce Human Sense - The Electronic Nose Artificial Olfaction Technology - page 139

Euclidean distance, (2) defining a clustering criterion to be optimized, typically
based on within- and between-cluster structure (e.g., elongated, compact or
topologically-ordered clusters), and (3) defining a search algorithm to find a
''good'' assignment of examples to clusters, since exhaustive enumeration of all
possible clustering is clearly impractical [ 21 ].

- Hierarchical Clustering: These algorithms are capable of generating a multi-
level clustering using a tree structure known as a dendrogram. These

dendrograms can be constructed in a bottom-up or top-down method.
- Self-Organizing Maps: SOMs are connectionist techniques capable of gen-
erating topology-preserving clustering [ 22 ]. An SOM is a network of clusters
(or neurons) arranged in a lattice structure, typically two-dimensional. The
behavior of SOMs result from the synergy of three processes: competition,
cooperation, and adaptation [ 23 ]. First, all neurons in the lattice enter a
competition for each incoming example. The closest neuron in feature space is
selected as a winner and becomes activated. SOMs have very interesting
properties for data visualization but mapping onto the SOM manifold can be
tricky if the structure of the data is inherently high dimensional.
- C means: C-means is a clustering algorithm that generates a single-level
partition of the dataset. Starting from an initial clustering (e.g., a random
assignment of examples to clusters), C-means iteratively re-computes the
sample mean of each cluster and reassigns each example to the cluster with
the closest mean. The basic C-means algorithm requires a pre-specified
number of clusters, heuristic procedures [ 24 ] can be employed to automati-
cally determine an appropriate number of clusters.

References

1. D.T. Win, The electronic nose—a big part of our future. AU. J. T. 9(1), 1-8 (2005)
2. R. Haddad, A. Medhanie, Y. Roth, D. Harel, N. Sobel, Predicting odour pleasantness with an
electronic nose. PLoS Comput. Biol. 6(4), e1000740 (2010). doi: 10.1371/journal.pcbi.

1000740
3. M. Bernardine Dias, Investigating the viability Of MEMS vapor sensors for detecting land
mines, CMU-RI-TR-00-24, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213,
October 2000
4. J.W. Gardner, P.N. Bartlett, Electronic Noses. Principles and Applications (Oxford

University Press, Oxford, 1999), pp. 221-245
5. AD Wilson et al., Applications and advances in electronic-nose technologies. Sensors 9,
5099-5148 (2009)
6. M. Ghasemi-Varnamkhasti et al., J. Food Eng. 100, 377-387 (2010)
7. K. Toko, Electronic tongue. Biosens. Bioelectron. 13, 701-709 (1998)
8. C. Di Natale, E. Mazzone, A. Mantini, A. Bearzotti, A. D'Amico, A.V. Legin, A.M.
Rudnitskaya, Y.G. Vlasov, Electronic tongue distinguishes different mineral waters. Alta
Frequenza Rivista Di Elettronica 11(2), 88-90 (1999)