Abstract
A recent novel approach to the visualisation and analysis of datasets, and one which is particularly applicable to those of a high dimension, is discussed in the context of real applications. A feed-forward neural network is utilised to effect a topographic, structure-preserving, dimension-reducing transformation of the data, with an additional facility to incorporate different degrees of associated subjective information. The properties of this transformation are illustrated on synthetic and real datasets, including the 1992 UK Research Assessment Exercise for funding in higher education. The method is compared and contrasted to established techniques for feature extraction, and related to topographic mappings, the Sammon projection and the statistical field of multidimensional scaling.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
Fukunaga, K. Introduction to Statistical Pattern Recognition. Academic Press, New York, second edition, 1990
Koontz WLG, Fukunaga K. A nonlinear feature extraction algorithm using distance transformation. IEEE Trans Comput 1972; C-21(1): 56–63
Cox TF, Ferry G. Discriminant analysis using non-metric multidimensional scaling. Patt Recogn 1993; 26(1): 145–153
Webb AR. Multidimensional scaling by iterative majorisation using radial basis functions. Patt Recogn 1995; 28(5): 753–759
Lowe D. Novel ‘topographic’ nonlinear feature extraction using radial basis functions for concentration coding in the ‘artificial nose’. In: 3rd IEE International Conference on Artificial Neural Networks, IEE, London, 1993
Press WH, Teukolsky SA, Vetterling WT, Flannery BP. Numerical Recipes in C. Cambridge University Press, second edition, 1992
Mardia KV, Kent JT, Bibby JM. Multivariate Analysis: Probability and Mathematical Statistics. Academic Press, London, 1979
Oja E. Neural networks, principal components, and subspaces. Int Neural Systems 1989; 1: 61–68
Saund E. Dimensionality-reduction using connectionist networks. IEEE Trans Patt Analysis Machine Intell 1989; 11(3): 304–315
Kramer MA. Nonlinear principal component analysis using autoassociative neural networks. AIChE J 1991; 37(2): 233–243, February
Hastie T, Stuetzle W. Principal curves. J Am Statist Assoc 1989; 84: 502–516
Fisher RA. The use of multiple measurements on taxonomic problems. Ann Eugenics 1936; 7: 179–188
Gallinari P, Thiria S, Badran B, Fogelman-Soulié F. On the relations between discriminant analysis and multilayer perceptrons. Neural Networks 1991; 4: 349–360
Webb AR, Lowe D. The optimised internal representation of multilayer classifier networks performs nonlinear discriminant analysis. Neural Networks 1990; 3(4): 367–375
Willshaw DJ, von der Malsburg C. How patterned neural connections can be set up by self-organisation. Proc Roy Soc Lond B 1976; 194: 431–445
Kohonen T. Self-Organisation and Associative Memory. Springer-Verlag, New York, third edition, 1989
Li X, Gasteiger J, Zupan J. On the topology distortion in self-organising feature maps. Biol Cybernetics 1993; 70: 189–198
Sammon JW, Jr. A nonlinear mapping for data structure analysis. IEEE Trans Comput 1969; C-18(5): 401–409
Davison ML. Multidimensional Scaling. John Wiley & Sons, New York, 1983
Young MP. Objective analysis of the topological organization of the primate cortical visual system. Nature 1992; 358: 152–155, July
Goodhill GJ, Simmen MW, Willshaw DJ. An evaluation of the use of multidimensional scaling for understanding brain connectivity. Philos Trans Roy Soc Lond Series B 1995; 348: 265–280
Torgerson WS. Multidimensional scaling: I. theory and method. Psychometrika 1952; 17: 401–419
Shepard RN. The analysis of proximities: Multidimensional scaling with an unknown distance function, I. Psychometrika 1962; 27(2): 125–140
Shepard RN. The analysis of proximities: Multidimensional scaling with an unknown distance function, II. Psychometrika 1962; 27(3): 219–246
Kruskal JB. Multidimensional scaling by optimising goodness of fit to a nonmetric hypothesis. Psychometrika 1964; 29(1): 1–27
Kruskal JB. Nonmetric multidimensional scaling: a numerical method. Psychometrika 1964; 29(2): 115–129
Kruskal JB. Comments on ‘a nonlinear mapping for data structure analysis’. IEEE Trans Comput 1971; C20: 1614
Jain AK, Jianchang Mao. Artificial neural network for nonlinear projection of multivariate data. In: IJCNN Int Joint Conf Neural Networks, vol 3, 1992; 335–340
Jianchang Mao, Jain AK. Artificial neural networks for feature extraction and multivariate data projection. IEEE Trans Neural Networks 1995; 6(2): 296–317
Lowe D, Tipping M. A novel neural network technique for exploratory data analysis. In: Proc ICANN '95 (Scientific Conf) vol 1, 1995; 339–344
Webb AR. Non-metric multidimensional scaling using feed-forward networks. CSE1 Research Note 192, Defence Research Agency, Malvern, UK, June 1992
Taylor J. Measuring research performance in business and management studies in the United Kingdom: The 1992 research assessment exercise. Br J Management 1994; 5: 275–288
Taylor J. A statistical analysis of the 1992 Research Assessment Exercise. J Roy Statist Soc A 1995; 158(2): 241–261
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Lowe, D., Tipping, M. Feed-forward neural networks and topographic mappings for exploratory data analysis. Neural Comput & Applic 4, 83–95 (1996). https://doi.org/10.1007/BF01413744
Issue Date:
DOI: https://doi.org/10.1007/BF01413744