Forecasting with artificial neural networks: the state of the art

Artificial Neural Networks (ANNs) are being used increasingly to predict and forecast water resources variables. In this paper, the steps that should be followed in the development of such models are outlined. These include the choice of performance criteria, the division and pre-processing of the available data, the determination of appropriate model inputs and network architecture, optimisation of the connection weights (training) and model validation. The options available to modellers at each of these steps are discussed and the issues that should be considered are highlighted. A review of 43 papers dealing with the use of neural network models for the prediction and forecasting of water resources variables is undertaken in terms of the modelling process adopted. In all but two of the papers reviewed, feedforward networks are used. The vast majority of these networks are trained using the backpropagation algorithm. Issues in relation to the optimal division of the available data, data pre-processing and the choice of appropriate model inputs are seldom considered. In addition, the process of choosing appropriate stopping criteria and optimising network geometry and internal network parameters is generally described poorly or carried out inadequately. All of the above factors can result in non-optimal model performance and an inability to draw meaningful comparisons between different models. Future research efforts should be directed towards the development of guidelines which assist with the development of ANN models and the choice of when ANNs should be used in preference to alternative approaches, the assessment of methods for extracting the knowledge that is contained in the connection weights of trained ANNs and the incorporation of uncertainty into ANN models.

Neural networks for the prediction and forecasting of water resources variables: a review of modelling issues and applications

This paper deals with the application of a novel neural network technique, support vector machine (SVM), in financial time series forecasting. The objective of this paper is to examine the feasibility of SVM in financial time series forecasting by comparing it with a multi-layer back-propagation (BP) neural network. Five real futures contracts that are collated from the Chicago Mercantile Market are used as the data sets. The experiment shows that SVM outperforms the BP neural network based on the criteria of normalized mean square error (NMSE), mean absolute error (MAE), directional symmetry (DS) and weighted directional symmetry (WDS). Since there is no structured way to choose the free parameters of SVMs, the variability in performance with respect to the free parameters is investigated in this study. Analysis of the experimental results proved that it is advantageous to apply SVMs to forecast financial time series.

Application of support vector machines in financial time series forecasting

A novel type of learning machine called support vector machine (SVM) has been receiving increasing interest in areas ranging from its original application in pattern recognition to other applications such as regression estimation due to its remarkable generalization performance. This paper deals with the application of SVM in financial time series forecasting. The feasibility of applying SVM in financial forecasting is first examined by comparing it with the multilayer back-propagation (BP) neural network and the regularized radial basis function (RBF) neural network. The variability in performance of SVM with respect to the free parameters is investigated experimentally. Adaptive parameters are then proposed by incorporating the nonstationarity of financial time series into SVM. Five real futures contracts collated from the Chicago Mercantile Market are used as the data sets. The simulation shows that among the three methods, SVM outperforms the BP neural network in financial forecasting, and there are comparable generalization performance between SVM and the regularized RBF neural network. Furthermore, the free parameters of SVM have a great effect on the generalization performance. SVM with adaptive parameters can both achieve higher generalization performance and use fewer support vectors than the standard SVM in financial forecasting.

/pdf/support-vector-machine-with-adaptive-parameters-in-financial-x0iweyp59q.pdf

Support vector machine with adaptive parameters in financial time series forecasting

Role playing can be used to forecast decisions, such as “how will our competitors respond if we lower our prices?” In role playing, an administrator asks people to play roles and uses their “decisions” as forecasts. Such an exercise can produce a realistic simulation of the interactions among conflicting groups. The role play should match the actual situation in key respects, such as that role players should be somewhat similar to those being represented in the actual situations, and roleplayers should read instructions for their roles before reading about the situation. Role playing is most effective for predictions when two conflicting parties respond to large changes. A review of the evidence showed that role playing was effective in matching results for seven of eight experiments. In five actual situations, role playing was correct for 56 percent of 143 predictions, while unaided expert opinions were correct for 16 percent of 172 predictions. Role playing has also been used successfully to forecast outcomes in three studies. Successful uses of role playing have been claimed in the military, law, and business.

Principles of forecasting

Designing a neural network for forecasting financial and economic time series

A comparison of artificial neural network and time series models for forecasting commodity prices

/pdf/forecasting-futures-trading-volume-using-neural-networks-239lsotz1f.pdf

Forecasting futures trading volume using neural networks

Neural networks are nonlinear mapping structures based on the study of the human brain. They have been shown to be universal and highly flexible junction approximators to any data-generating process. Therefore, they are powerful models for forecasting purposes, especially when the underlying data-generating processes are unknown. However, the appropriate design of the network's architecture and learning rules are crucial for obtaining satisfactory results. This study discusses the scope and limitations of neural networks for forecasting problems and provides an example by designing a neural network for forecasting. It is argued that statistical theory can offer some suggestions for designing an optimal network architecture. An example comparing a neural network and ARIM model for forecasting weekly corn prices 1974 through 1993 is provided. Results show the neural network model to be more accurate than the ARIMA.



Les reseaux neuraux sont des structures de cartographic non lineaires bases sur l'etude du cerveau humain. Ils se sont reveles etre des approximateurs de fonction universele et d'une grande souplesse, applicables a tout processus de generation de donnees. Pour cette raison, ce sont depuissants modeles de prediction, surtout lorsqu‘on ignore les processus de generation des donnees sous-jacents. Toutefois, le concept architectural du reseau et le regles d'apprentissage doivent convenir aux resultats que l'on cherche a obtenir. La presenteetude porte sur l'envergure et sur les limites des reseaux neuraux pour la prediction des problemes etfournit un exemple de reseau neural. L'argumentation suppose que la theorie statistique peut offrir certaines suggestions pour la conception d'une architecture de reseau optimale. Un exemple sert a comparer un reseau neural et un modele ARIMA pour la production des prix hebdomadaires du mats de 1974 a 1993. Les resultats montrent que le modele a reseau neural est plus juste que le modele ARIMA.

Neural Networks for Forecasting: An Introduction

Neural networks are a relatively new computer artificial intelligence method which attempt to mimic the brain's problem solving process and can be used for predicting nonlinear economic time series. Neural networks are used to look for patterns in data, learn these patterns, and then classify new patterns and make forecasts. Feedforward neural networks pass the data forward from input to output, while recurrent networks have a feedback loop where data can be fed back into the input at some point before it is fed forward again for further processing and final output. Some have argued that since time series data may have autocorrelation or time dependence, the recurrent neural network models which take advantage of time dependence may be useful. Feedforward and recurrent neural networks are used for comparison in forecasting the Japanese yen/US dollar exchange rate. A traditional ARIMA model is used as a benchmark for comparison with the neural network models.

Iebeling Kaastra

Papers

Designing a neural network for forecasting financial and economic time series

A comparison of artificial neural network and time series models for forecasting commodity prices

Forecasting futures trading volume using neural networks

Neural Networks for Forecasting: An Introduction

Feedforward versus recurrent neural networks for forecasting monthly japanese yen exchange rates