Artificial Neural Networks. A New Approach to Modelling Interregional Telecommunication Flows

Fischer, Manfred M. and Gopal, Sucharita (1994): Artificial Neural Networks. A New Approach to Modelling Interregional Telecommunication Flows. Published in: Journal of Regional Science , Vol. 34, No. 4 (1994): pp. 503-527.

Abstract

During the last thirty years there has been much research effort in regional science devoted to modelling interactions over geographic space. Theoretical approaches for studying these phenomena have been modified considerably. This paper suggests a 'new modelling approach, based upon a general nested sigmoid neural network model. Its feasibility is illustrated in the context of modelling interregional telecommunication traffic in Austria and its performance is evaluated in comparison with the classical regression approach of the gravity type. The application of this neural network approach may be viewed as a three-stage process. The first stage refers to the identification of an appropriate network from the family of two-layered feedforward networks with 3 input nodes, one layer of (sigmoidal) intermediate nodes and one (sigmoidal) output node (logistic activation function). There is no general procedure to address this problem. We solved this issue experimentally. The input-output dimensions have been chosen in order to make the comparison with the gravity model as close as possible. The second stage involves the estimation of the network parameters of the selected neural network model. This is perlormed via the adaptive setting of the network parameters (training, estimation) by means of the application of a least mean squared error goal and the error back propagating technique, a recursive learning procedure using a gradient search to minimize the error goal. Particular emphasis is laid on the sensitivity of the network perlormance to the choice of the initial network parameters as well as on the problem of overlitting. The final stage of applying the neural network approach refers to the testing of the interregional teletraffic flows predicted. Prediction quality is analysed by means of two perlormance measures, average relative variance and the coefficient of determination, as well as by the use of residual analysis. The analysis shows that the neural network model approach outperlorms the classical regression approach to modelling telecommunication traffic in Austria.

• Item Type: MPRA Paper
• Language: English
• Keywords: n.a.
• Subjects: C - Mathematical and Quantitative Methods > C4 - Econometric and Statistical Methods: Special Topics > C45 - Neural Networks and Related Topics

References

Amari, Shun-lchi (1990): Mathematical foundations of neurocomputing, Proceedings of the IEEE 78, pp. 1443-1463.

Barron, Andrew R. (1993): Universal approximation bounds for superpositions of a sigmoidal function, IEEE Transactions on Information Theory 39, pp. 930-945.

Bishop, Chris (1991 ): Improving the generalization properties of radial basis function neural networks, Neural Computation 3, pp. 579-588.

Cybenko, G. (1989): Approximation by superpositions of sigmoidal function, Mathematics of Control, Signals, and Systems 2, pp. 303-314.

Fischer, Manfred M. (1993): Travel demand, in Polak, Jacob and Heertje, Arnold (eds.): European Transportation Economics, pp. 6-32. Oxford (UK) and Cambridge (MA): Basil Blackwell.

Fischer, Manfred M. and Gopal, Sucharita (1993): Neurocomputing - a new paradigm for geographic information processing, Environment and Planning A 25, pp. 757-760.

Fischer, Manfred M., Essetzbichler, Jurgen, Gassler, Helmut and Trichtl, Gerhard (1992): Interregional and international telephone communication. Aggregate traffic models and empirical evidence tor Austria, Sistemi Urbani (in press).

Fotheringham, Stewart A., and O'Kelly, Morton E. (1989): Spatial Interaction Models: Formulations and Applications. Dordrecht, Boston and Lodon: Kluwer.

Funahashi, Ken-lchi (1989): On the approximate realization of continous mappings by neural networks, Neural Networks 2, pp. 183-192.

Gillespie, Andrew and Williams, Howard (1988): Telecommunications and the reconstruction of regional comparative advantage, Environment and Planning A 20, pp. 1311-1321.

Guldmann, Jean-Michel (1992): Modeling residential and business telecommunication flows: A regional point-to-point approach, Geographical Analysis 24, pp. 121-141.

Gyer, Maurice S. (1992): Adjuncts and alternatives to neural networks for supervised classification, IEEE Transactions on Systems, Man, and Cybernetics 22, pp. 35-47.

Hecht-Nielsen, Robert (1990): Neurocomputing. Reading (MA): Addison-Wesley. Hornik, Kurt, Stinchcombe, Maxwell and White, Halbert (1989): Multilayer feedforward networks are universal approximators, Neural Network 2, pp. 359-366.

Levin, Esther, Tishby, Naftali, and Solla, Sara A. (1990): A statistical approach to learning and generalization in layered neural networks, Proceedings of the IEEE 78, pp. 1568-1575.

Openshaw, Stan (1992): Modelling spatial interaction using a neural net, in Fischer, Manfred M. and Nijkamp, Peter (eds): Geographic Information Systems, Spatial Modelling and Policy Evaluation, pp. 147-164. Berlin: Springer.

Pacey, P. L. (1983): Long distance demand: A point-to-point model, Southern Economic Journal 49, pp. 1094-1107.

Rietveld, Piet and Janssen, Leon (1990): Telephone calls and communication barriers. The case of the Netherlands, The Annals of Regional Science 24, pp. 307-318. Rossera, Fabio (1990): Discontinuities and barriers in communications. The case of Swiss communities of different language, The Annals of Regional Science 24, pp. 319-336.

Rummelhart, David E., Hinton, Geoffrey E. and Williams, Ronald J. (1986): Learning internal representations by error propagation; in Rummelhart, David E., McClelland, James L. and the PDP Research Group (eds.) (1986): Parallel Distributed Processing. Explorations in the Microstructures of Cognition. Volume 1: Foundations, pp. 318-362. Cambridge (Ma.) and London (England): The MIT Press.

Salomon, llan (1986): Telecommunications and travel relationships: A review, Transportation Research 3, pp. 223-238.

Shah, Samir, Palmieri, Francesco and Datum, Michael (1992): Optimal filtering algorithms for fast learning in feedforward neural networks, Neural Networks 5, pp. 779-787.

Weigend, Andreas S., Rumelhart, David E. and Huberman, Bernardo A. (1991 ): Back-propagation, weightelimination and time series prediction, in Touretzki, David S., Elman, Jeffrey L., Sejnowski, Terrence J., and Hinton, Geoffrey E. (eds.): Connectionist Models. Proceedings of the 1990 Summer School, pp.105-116. San Mateo (CA), Morgan Kaufmann Publishers.

White, Halbert (1989): Some asymptotic results for learning in single hidden-layer feedforward network models, Journal of American Statistical Association 84, pp. 1003-1013.

Widrow, Bernard , and Lehr, Michael A. (1990): 30 years of adaptive neural networks: Perceptron, Madaline, and Backpropagation, Proceedings of the IEEE 78, pp. 1415-1447.