Predicting the Future: A Connectionist Approach

Abstract

We investigate the effectiveness of connectionist architectures for predicting the future behavior of nonlinear dynamical systems. We focus on real-world time series of limited record length. Two examples are analyzed: the benchmark sunspot series and chaotic data from a computational ecosystem. The problem of overfitting, particularly serious for short records of noisy data, is addressed both by using the statistical method of validation and by adding a complexity term to the cost function ("back-propagation with weight-elimination"). The dimension of the dynamics underlying the time series, its Liapunov coefficient, and its nonlinearity can be determined via the network. We also show why sigmoid units are superior in performance to radial basis functions for high-dimensional input spaces. Furthermore, since the ultimate goal is accuracy in the prediction, we find that sigmoid networks trained with the weight-elimination algorithm outperform traditional nonlinear statistical approaches.

Prediction of Molar Absorptivities of Color Reagents during Their Color Reactions with Cerium Using Multiple Regression Analysis and Neural Network

“…This problem has been overcome by two methods in the past: by using the method of weight elimination and by using an external cross-validation set. 20 In this experiment, we chose the test set as a monitor to control the training processing. In this way, the performance of the neural network can be enchanted markedly.…”

Study on Decision-Making Model of Bid Evaluation for Engineering Projects Based on BP Neural Network and Order Relation Analysis

“…Back propagation neural network [11,12] is a multi-layer network model of feedback type proposed by the famous researcher Rumelhart [13] in 1985, which is one of the most widely used artificial neural network models. It consists of input layer, hidden layer and output layer.…”

Neural Network For Modeling Nonlinear Time Series: A New Approach

“…It has been proved that this class of networks can approximate continuous functions at any arbitrary accuracy. A complete description of a Neural Networks theory and the application of neural networks to the problem of nonlinear system identification and prediction can be found in [5], [6]. The non linear coefficients ϕ i (X t ) and γ j (X t ) in (1) can be approximated by a sub-THNN, then:…”