ARTIFICIAL NEURAL NETWORKS: A NEW APPROACH TO MODELING INTERREGIONAL TELECOMMUNICATION FLOWS*

ABSTRACT. During the last thirty years there has been much research effort in regional science devoted to modeling interactions over geographic space. Theoretical approaches for studying these phenomena have been modified considerably. This paper suggests a new modeling approach, based upon a general nested sigmoid neural network model. Its feasibility is illustrated in the context of modeling 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 performed 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 performance to the choice of the initial network parameters, as well as on the problem of overfitting. The final stage of applying the neural network approach refers to the testing of the interregional teletraffic flows predicted. Prediction quality is analyzed by means of two performance 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 outperforms the classical regression approach to modeling telecommunication traffic in Austria.     

Tuned Mass Damper Positioning Effects on the Seismic Response of a Soil-MDOF-Structure System

Tuned mass dampers (TMDs) are effective structural vibration control devices. However, very little research is available on the experimental investigation of TMDs and their performance in systems undergoing dynamic soil-structure interaction. Geotechnical centrifuge tests are conducted to investigate story positioning effects of single and multiple TMDs in a soil-MDOF-structure system. The criteria for optimal story positioning will be established, and it is shown that story positioning influences TMD performance more than the number of TMDs used. Non-optimal story positioning was found to have the potential of reducing damping efficiency, amplifying peak structural response, and inducing lengthier high-intensity motion.     

Application of Artificial Neural Networks in Climatology: A Case Study of Sunspot Prediction and Solar Climate Trends

Global temperature trends on time scales of years to centuries have recently been shown to be related to volcanic aerosols, carbon dioxide levels, and solar activity. The most visible and well-studied indicators of solar variability are dark areas or “sunspots” on the surface of the Sun, with sunspot numbers directly related to the level of solar activity. Prediction of sunspot numbers in advance of the actual event has proven problematic with most methods failing due to nonlinearities in solar activity. An approach using the generation of a feedforward neural network may resolve some of the difficulties inherent in currently utilized statistical and precursor approaches since feedforward networks offer a useful and practical method of approximating nonlinear relations and their derivatives without knowing the actual underlying nonlinear function. In this paper, we show some preliminary findings in using feedforward neural networks for the prediction of peak sunspot cycle amplitude and discuss the climatic implications of the findings.     

Learning in Single Hidden-Layer Feedforward Network Models: Backpropagation in a Spatial Interaction Modeling Context

Learning in neural networks has attracted considerable interest in recent years. Our focus is on learning in single hidden-layer feedforward networks which is posed as a search in the network parameter space for a network that minimizes an additive error function of statistically independent examples. We review first the class of single hidden-layer feedforward networks and characterize the learning process in such networks from a statistical point of view. Then we describe the backpropagation procedure, the leading case of gradient descent learning algorithms for the class of networks considered here, as well as an efficient heuristic modification. Finally, we analyze the applicability of these learning methods to the problem of predicting interregional telecommunication flows. Particular emphasis is laid on the engineering judgment, first, in choosing appropriate values for the tunable parameters, second, on the decision whether to train the network by epoch or by pattern (random approximation), and, third, on the overfitting problem. In addition, the analysis shows that the neural network model whether using either epoch-based or pattern-based stochastic approximation outperforms the classical regression approach to modeling telecommunication flows.