Linear versus neural network forecasts for European industrial production series

Heravi, Saeed; Osborn, Denise R.; Birchenhall, Chris

doi:10.1016/s0169-2070(03)00062-1

Cited by 91 publications

(49 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…One major application areas of NNs is forecasting, and the NNs techniques have been used as to solve many forecasting problems ( [33], [36], [39], [38].…”

Section: Neural Networkmentioning

confidence: 99%

Data Mining and Neural Networks: The Impact of Data Representation

Siraj¹,

Omer²,

Hasan³

2012

Advances in Data Mining Knowledge Discovery and Applications

View full text Add to dashboard Cite

“…One major application areas of NNs is forecasting, and the NNs techniques have been used as to solve many forecasting problems ( [33], [36], [39], [38].…”

Section: Neural Networkmentioning

confidence: 99%

Data Mining and Neural Networks: The Impact of Data Representation

Siraj¹,

Omer²,

Hasan³

2012

Advances in Data Mining Knowledge Discovery and Applications

View full text Add to dashboard Cite

“…23 Auch im Gebiet der makroökonomischen Zeitreihen liegen mittlerweile viele Studien vor, so dass auch hier nur eine Auswahl gemacht werden kann: Swanson und White (1997b), Swanson und White (1997a), Stock und Watson (1999), Moshiri und Cameron (2000), Balkin und Ord (2000), Rech (2002), Heravi et al (2004), Teräsvirta et al (2005). Unter diesen Artikeln sind die beiden letztgenannten für die vorliegene Untersuchung wichtig, da sie besonders umfassend sind undähnliche Modellierungsansätze verfolgen wie die in diesem Kapitel beschriebenen.…”

Section: Neuronale Netze Zur Zeitreihenmodellierung Und -Prognose Imüunclassified

Prognose makroökonomischer Zeitreihen: Ein Vergleich linearer Modelle mit neuronalen Netzen

Köller¹

2014

View full text Add to dashboard Cite

This thesis examines the potential of neural networks (NN) for forecasting of macroeconomic time series in comparison with linear models. The emphasis is on the autoregressive neural network (ARNN) model, which can be seen as a generalisation of the conventional autoregressive (AR) model where the non-linear part is implemented by a neural network of the feedforward type.Among the properties of macroeconomic time series that motivate various adaptations and enhancements both of linear and of neural network methodology are the following: high stochasticity, short length, non-stationary and seasonal behaviour and the fact that the existence of non-linearity is not clearly known a priori. The development and evaluation of the methods is based on simulated time series with known non-linear properties and on two macroeconomic time series: Austrian unemployment rate and Austrian industrial production index. Both series are in monthly observations and are seasonally unadjusted.After the introduction, the first main chapter of the thesis summarizes the basic concepts of linear univariate time series analysis and applies the linear methodology to the two selected macroeconomic time series. Particular attention is devoted to the problem of finding the appropriate differencing filter -conventional differences, seasonal differences or both -as a way to induce stationarity. Amongst other methods unit root tests are employed. The used linear models are the autoregressive (AR) and the autoregressive moving average (ARMA) model. These are augmented by a model part for deterministic seasonality. The models estimated on the two time series have rather high model order and, despite using a sparse modelling technique, comprise relatively many coefficients.The second main chapter treats various theoretic aspects and types of nonlinearity and carries out a sequence of hypothesis tests for non-linearity, in order to be able to rate the possible benefits of applying neural network methods to the two selected time series. The results of these tests indicate the presence of non-linear structure of the additive type, which can be exploited for better forecasting with ARNN models. However, these results have to be interpreted with care, as they might bei compromised by non-linearity of the multiplicative type, residual linear structure and possible structural breaks.The third main chapter introduces the ARNN model and develops methods for its estimation and specification. The specification contains a model part of determistic seasonality, includes a linear part and allows for sparse specification of coefficients. The generalisation capability of the ARNN model is to be ensured by four alternative model building approaches: the statistical-parametric approach with hypothesis testing and pruning, the classical approach with early stopping, the approach with regularisation and the Bayesian evidence framework. For each one of these approaches modifications and additional heuristics are contributed that seem appropriate in view of the a...

show abstract

“…Para abordar este problema se han desarrollado numerosas técnicas que abarcan modelos estadísticos tradicionales tales como los ARIMA y las funciones de transferencia [2], hasta modelos no lineales [3]. Particularmente, los modelos de redes neuronales artificiales han sido de especial utilidad cuando la serie estudiada presenta características no lineales; en [4] se presenta una revisión general, y aplicaciones específicas en [5][6][7][8][9][10][11][12][13].…”

Section: Introductionunclassified

Predicción De Series Temporales Usando Máquinas De Vectores De Soporte

Velásquez

Morales

Franco

2010

Ingeniare. Rev. chil. ing.

View full text Add to dashboard Cite

RESUMENLa predicción de series de tiempo es un importante problema de investigación debido a sus implicaciones en ingeniería, economía, finanzas y ciencias sociales. Un importante tópico de esta problemática es el desarrollo de nuevos modelos y su comparación con aproximaciones previas en términos de la precisión del pronóstico. Recientemente, las máquinas de vectores de soporte (SVM) han sido usadas para la predicción de series de tiempo, pero las experiencias reportadas son limitadas y hay algunos problemas relacionados con su especificación. El objetivo de este artículo es proponer una técnica novedosa para estimar algunas constantes en las SVM que usualmente son fijadas en forma empírica por el modelador. La técnica propuesta es usada para estimar varias SVM con el fin de pronosticar cinco series benchmark; los resultados obtenidos son comparados con las estadísticas reportadas en otros artículos. La metodología propuesta permite obtener SVM competitivas para las series pronosticadas en comparación con los resultados obtenidos usando otros modelos más tradicionales.Palabras clave: Series de tiempo, predicción, redes neuronales artificiales, máquinas de vectores de soporte. Particularmente, los perceptrones multicapa (MLP, por sus siglas en inglés) han sido ampliamente utilizados para la predicción de series no lineales, debido a que pueden aproximar cualquier función continua definida en un dominio compacto [14][15][16]. Sin embargo, su especificación se basa fundamentalmente en criterios heurísticos y juicio experto del modelador [17,18], de tal manera, que dicho proceso se basa en un conjunto de pasos críticos que ABSTRACT Time series prediction is an important research problem due to its implications in engineering

show abstract

Linear versus neural network forecasts for European industrial production series

Cited by 91 publications

References 22 publications

Data Mining and Neural Networks: The Impact of Data Representation

Data Mining and Neural Networks: The Impact of Data Representation

Prognose makroökonomischer Zeitreihen: Ein Vergleich linearer Modelle mit neuronalen Netzen

Predicción De Series Temporales Usando Máquinas De Vectores De Soporte

Contact Info

Product

Resources

About