All models are allowed to evolve through time, and our analysis focuses on model selection and performance. Macroeconomic variables are forecast in a real-time scenario using a variety of flexible specification, fixed specification, linear, and nonlinear econometric models. In the context of real-time forecasts, flexible specification models (including linear autoregressive models with exogenous variables and nonlinear artificial neural networks) appear to offer a useful and viable alternative to less flexible fixed specification linear models for a subset of the economic variables which we examine, particularly at forecast horizons greater than 1-step ahead.
One reason for this result is that the economy is evolving (rather slowly) over time. This feature cannot easily be captured by fixed specification linear models, however, and manifests itself in the form of evolving coefficient estimates. We also provide additional evidence supporting the claim that models which ‘win’ based on one model selection criterion (say a squared error measure) do not necessarily win when an alternative selection criterion is used (say a confusion rate measure), thus highlighting the importance of the particular cost function which is used by forecasters and ‘end-users’ to evaluate their models(1). A wide variety of different model selection criteria and statistical tests are used to illustrate our findings.
The examination of the neural networks usage as an alternative to classical statistical techniques for forecasting within the framework of the APT (arbitrage pricing theory) model for stock ranking showed that neural networks outperform these statistical techniques in forecasting accuracy terms(2), and give better model fitness in-sample by one order of magnitude.
Kai Chun Chiu and Lei Xuviewed as a promising application of neural networks, financial time series forecasting based on recently developed Temporal Factor Analysis (TFA) model mainly targeted at further study of the Arbitrage Pricing Theory (APT).They found that there’s a potential application TFA in the prediction of stock price and index, aiming to illustrate the superiority of using the APT-based Gaussian TFA model as compared to three conventional approaches which are not financial model-based (3). We tend to consider that demonstrated N-ENRBF approachin general more difficult to anticipate in cases when nonstationary signals, such as the those referred to as random fluctuations in prices are taken place.
Summary.The intervals for the network parameter values for which these performance figures are statistically stable. Neural networks have been criticized for not being able to provide an explanation of how they interact with their environment and how they reach an outcome. We show that by using sensitivity analysis, neural networks can provide a reasonable explanation of their predictive behavior and can model their environment more convincingly than regression models.
References
1. Apostolos Nicholas R., Achileas Z., Gavin F. Neural Networks, Volume 7, Issue 2, 1994, Pages 375–388.
2. Guoqiang Zhang, B. Eddy Patuwo, Michael Y. HuForecasting with artificial neural networks:: The state of the art. International Journal of Forecasting, Volume 14, Issue 1, 1 March 1998, Pages 35–62.
3. K. Chiu, L. Xu, (2002) “Stock price and index forecasting by arbitrage pricing theory-based Gaussian TFA learning”, Intelligent Data Engineering and Automated Learning – IDEAL 2002, LNCS 2412 pp.366-371, Springer Verlag.
|
Голосування 
Ця робота ліцензується відповідно до Creative Commons Attribution 4.0 International License
Знайшли помилку? Виділіть помилковий текст мишкою і натисніть Ctrl + Enter