A model selection approach to real-time macroeconomic forecasting using linear models and artificial neural networks

Article Abstract:

A study aimed at determining whether a class of adaptive prediction models or artificial neural networks is efficient in forecasting future values of nine macroeconomic variables adopted a model selection approach. Various out-of-sample forecast-based model selection criteria, including forecast error methods and forecast direction accuracy were utilized. Results obtained in real-time forecasting based on rolling window prediction measures reveal multivariate adaptive linear vector autoregression models usually surpass various models that are presented.

Economics, Research and Development in the Social Sciences and Humanities, Computer networks, Neural networks, Macroeconomics, Linear models (Statistics)

---

Encompassing the forecasts of U.S. trade balance models

Article Abstract:

An analysis describing the concept of forecast encompassing is presented. The concept refers to the absence of additional information in another economic model's forecasts. A test statistic is applied to another model's forecasts based on another model's error regression. A generalized statistic which is based on sets of dynamic nonlinear models with uncertain estimated coefficients is tested. It is shown that forecast application on US data involving trade balance of the generalized statistic indicates mis-specification.

Forecasts and trends, United States economic conditions, Balance of trade

---

Conditional forecasts in dynamic multivariate models

Article Abstract:

Bayesian methods are applied for the finite-sample distribution of conditional forecasts, an analytic approach which is useful for both vector autoregression and structural models.

Science & research, Econometrics & Model Building, Statistical Data Included, Methods, Econometrics, Bayesian statistical decision theory, Bayesian analysis, Multivariate analysis, Business models

Similar abstracts:

• Abstracts: A dynamic model selection procedure to forecast using multi-process models. Robust modelling of ARCH models
• Abstracts: Seasonal cointegration in macroeconomic systems: case studies for small and large European countries. Lifetime and annual marginal costs of redistribution in England, Sweden, and the United States
• Abstracts: Cooperative versus noncooperative behavior: the case of agricultural research. The effects of R&D, foreign technology purchase, and domestic and international spillovers on productivity in Indian firms
• Abstracts: Implications of ownership, regulation and market structure for performance: evidence from the U.S. electric utility industry before and after the new deal
• Abstracts: Economic factors and the stock market: a new perspective. Forecasting volatility in commodity markets. Assessing inefficiency in the S&P 500 futures market

This website is not affiliated with document authors or copyright owners. This page is provided for informational purposes only. Unintentional errors are possible.