Order series method for forecasting non‐Gaussian time series

A new forecasting non‐Gaussian time series method based on order series transformation properties has been proposed. The proposed method improves Yu's method without using Hermite polynomial expansion to process nonlinear instantaneous transformations and provides acceptable forecasting accuracy. Copyright © 2007 John Wiley & Sons, Ltd.     

A non‐Gaussian generalization of the Airline model for robust seasonal adjustment

In their seminal book Time Series Analysis: Forecasting and Control, Box and Jenkins (1976) introduce the Airline model, which is still routinely used for the modelling of economic seasonal time series. The Airline model is for a differenced time series (in levels and seasons) and constitutes a linear moving average of lagged Gaussian disturbances which depends on two coefficients and a fixed variance. In this paper a novel approach to seasonal adjustment is developed that is based on the Airline model and that accounts for outliers and breaks in time series. For this purpose we consider the canonical representation of the Airline model. It takes the model as a sum of trend, seasonal and irregular (unobserved) components which are uniquely identified as a result of the canonical decomposition. The resulting unobserved components time series model is extended by components that allow for outliers and breaks. When all components depend on Gaussian disturbances, the model can be cast in state space form and the Kalman filter can compute the exact log‐likelihood function. Related filtering and smoothing algorithms can be used to compute minimum mean squared error estimates of the unobserved components. However, the outlier and break components typically rely on heavy‐tailed densities such as the t or the mixture of normals. For this class of non‐Gaussian models, Monte Carlo simulation techniques will be used for estimation, signal extraction and seasonal adjustment. This robust approach to seasonal adjustment allows outliers to be accounted for, while keeping the underlying structures that are currently used to aid reporting of economic time series data. Copyright © 2006 John Wiley & Sons, Ltd.     

A non‐linear dynamic model for multiplicative seasonal‐trend decomposition

A non‐linear dynamic model is introduced for multiplicative seasonal time series that follows and extends the X‐11 paradigm where the observed time series is a product of trend, seasonal and irregular factors. A selection of standard seasonal and trend component models used in additive dynamic time series models are adapted for the multiplicative framework and a non‐linear filtering procedure is proposed. The results are illustrated and compared to X‐11 and log‐additive models using real data. In particular it is shown that the new procedures do not suffer from the trend bias present in log‐additive models. Copyright © 2002 John Wiley & Sons, Ltd.     

Bayesian Analysis of Vector ARMA Models using Gibbs Sampling

We present a methodology for estimation, prediction, and model assessment of vector autoregressive moving-average (VARMA) models in the Bayesian framework using Markov chain Monte Carlo algorithms. The sampling-based Bayesian framework for inference allows for the incorporation of parameter restrictions, such as stationarity restrictions or zero constraints, through appropriate prior specifications. It also facilitates extensive posterior and predictive analyses through the use of numerical summary statistics and graphical displays, such as box plots and density plots for estimated parameters. We present a method for computationally feasible evaluation of the joint posterior density of the model parameters using the exact likelihood function, and discuss the use of backcasting to approximate the exact likelihood function in certain cases. We also show how to incorporate indicator variables as additional parameters for use in coefficient selection. The sampling is facilitated through a Metropolis–Hastings algorithm. Graphical techniques based on predictive distributions are used for informal model assessment. The methods are illustrated using two data sets from business and economics. The first example consists of quarterly fixed investment, disposable income, and consumption rates for West Germany, which are known to have correlation and feedback relationships between series. The second example consists of monthly revenue data from seven different geographic areas of IBM. The revenue data exhibit seasonality, strong inter-regional dependence, and feedback relationships between certain regions.© 1997 John Wiley & Sons, Ltd.     

ARIMA model building and the time series analysis approach to forecasting

This paper reviews the approach to forecasting based on the construction of ARIMA time series models. Recent developments in this area are surveyed, and the approach is related to other forecasting methodologies.     

Bicorrelations and cross‐bicorrelations as non‐linearity tests and tools for exchange rate forecasting

This paper proposes and implements a new methodology for forecasting time series, based on bicorrelations and cross‐bicorrelations. It is shown that the forecasting technique arises as a natural extension of, and as a complement to, existing univariate and multivariate non‐linearity tests. The formulations are essentially modified autoregressive or vector autoregressive models respectively, which can be estimated using ordinary least squares. The techniques are applied to a set of high‐frequency exchange rate returns, and their out‐of‐sample forecasting performance is compared to that of other time series models. Copyright © 2001 John Wiley & Sons, Ltd.     

Unemployment variation over the business cycles: a comparison of forecasting models

Asymmetry has been well documented in the business cycle literature. The asymmetric business cycle suggests that major macroeconomic series, such as a country's unemployment rate, are non‐linear and, therefore, the use of linear models to explain their behaviour and forecast their future values may not be appropriate. Many researchers have focused on providing evidence for the non‐linearity in the unemployment series. Only recently have there been some developments in applying non‐linear models to estimate and forecast unemployment rates. A major concern of non‐linear modelling is the model specification problem; it is very hard to test all possible non‐linear specifications, and to select the most appropriate specification for a particular model. Artificial neural network (ANN) models provide a solution to the difficulty of forecasting unemployment over the asymmetric business cycle. ANN models are non‐linear, do not rely upon the classical regression assumptions, are capable of learning the structure of all kinds of patterns in a data set with a specified degree of accuracy, and can then use this structure to forecast future values of the data. In this paper, we apply two ANN models, a back‐propagation model and a generalized regression neural network model to estimate and forecast post‐war aggregate unemployment rates in the USA, Canada, UK, France and Japan. We compare the out‐of‐sample forecast results obtained by the ANN models with those obtained by several linear and non‐linear times series models currently used in the literature. It is shown that the artificial neural network models are able to forecast the unemployment series as well as, and in some cases better than, the other univariate econometrics time series models in our test. Copyright © 2004 John Wiley & Sons, Ltd.     

A forecasting procedure for nonlinear autoregressive time series models

Forecasting for nonlinear time series is an important topic in time series analysis. Existing numerical algorithms for multi‐step‐ahead forecasting ignore accuracy checking, alternative Monte Carlo methods are also computationally very demanding and their accuracy is difficult to control too. In this paper a numerical forecasting procedure for nonlinear autoregressive time series models is proposed. The forecasting procedure can be used to obtain approximate m‐step‐ahead predictive probability density functions, predictive distribution functions, predictive mean and variance, etc. for a range of nonlinear autoregressive time series models. Examples in the paper show that the forecasting procedure works very well both in terms of the accuracy of the results and in the ability to deal with different nonlinear autoregressive time series models. Copyright © 2005 John Wiley & Sons, Ltd.     

Modelling time series with season‐dependent autocorrelation structure

Time series with season‐dependent autocorrelation structure are commonly modelled using periodic autoregressive moving average (PARMA) processes. In most applications, the moving average terms are excluded for ease of estimation. We propose a new class of periodic unobserved component models (PUCM). Parameter estimates for PUCM are readily interpreted; the estimated coefficients correspond to variances of the measurement noise and of the error terms in unobserved components. We show that PUCM have correlation structure equivalent to that of a periodic integrated moving average (PIMA) process. Results from practical applications indicate that our models provide a natural framework for series with periodic autocorrelation structure both in terms of interpretability and forecasting accuracy. Copyright © 2008 John Wiley & Sons, Ltd.     

Short‐term forecasting of industrial electricity consumption in Brazil

This paper presents short‐term forecasting methods applied to electricity consumption in Brazil. The focus is on comparing the results obtained after using two distinct approaches: dynamic non‐linear models and econometric models. The first method, that we propose, is based on structural statistical models for multiple time series analysis and forecasting. It involves non‐observable components of locally linear trends for each individual series and a shared multiplicative seasonal component described by dynamic harmonics. The second method, adopted by the electricity power utilities in Brazil, consists of extrapolation of the past data and is based on statistical relations of simple or multiple regression type. To illustrate the proposed methodology, a numerical application is considered with real data. The data represents the monthly industrial electricity consumption in Brazil from the three main power utilities: Eletropaulo, Cemig and Light, situated at the major energy‐consuming states, Sao Paulo, Rio de Janeiro and Minas Gerais, respectively, in the Brazilian Southeast region. The chosen time period, January 1990 to September 1994, corresponds to an economically unstable period just before the beginning of the Brazilian Privatization Program. Implementation of the algorithms considered in this work was made via the statistical software S‐PLUS. Copyright © 1999 John Wiley & Sons, Ltd.     

A hybrid approach with step-size aggregation to forecasting hierarchical time series

Hierarchical time series arise in various fields such as manufacturing and services when the products or services can be hierarchically structured. “Top-down” and “bottom-up” forecasting approaches are often used for forecasting such hierarchical time series. In this paper, we develop a new hybrid approach (HA) with step-size aggregation for hierarchical time series forecasting. The new approach is a weighted average of the two classical approaches with the weights being optimally chosen for all the series at each level of the hierarchy to minimize the variance of the forecast errors. The independent selection of weights for all the series at each level of the hierarchy makes the HA inconsistent while aggregating suitably across the hierarchy. To address this issue, we introduce a step-size aggregate factor that represents the relationship between forecasts of the two consecutive levels of the hierarchy. The key advantage of the proposed HA is that it captures the structure of the hierarchy inherently due to the combination of the hierarchical approaches instead of independent forecasts of all the series at each level of the hierarchy. We demonstrate the performance of the new approach by applying it to the monthly data of ‘Industrial’ category of M3-Competition as well as on Pakistan energy consumption data.     

A state-dependent linear recurrent formula with application to time series with structural breaks

An underlying assumption in Multivariate Singular Spectrum Analysis (MSSA) is that the time series are governed by a linear recurrent continuation. However, in the presence of a structural break, multiple series can be transferred from one homogeneous state to another over a comparatively short time breaking this assumption. As a consequence, forecasting performance can degrade significantly. In this paper, we propose a state-dependent model to incorporate the movement of states in the linear recurrent formula called a State-Dependent Multivariate SSA (SD-MSSA) model. The proposed model is examined for its reliability in the presence of a structural break by conducting an empirical analysis covering both synthetic and real data. Comparison with standard MSSA, BVAR, VAR and VECM models shows the proposed model outperforms all three models significantly.     

Highest-density forecast regions for nonlinear and non-normal time series models

Forecast regions are a common way to summarize forecast accuracy. They usually consist of an interval symmetric about the forecast mean. However, symmetric intervals may not be appropriate forecast regions when the forecast density is not symmetric and unimodal. With many modern time series models, such as those which are non-linear or have non-normal errors, the forecast densities are often asymmetric or multimodal. The problem of obtaining forecast regions in such cases is considered and it is proposed that highest-density forecast regions be used. A graphical method for presenting the results is discussed.     

A fractal forecasting model for financial time series

Financial market time series exhibit high degrees of non‐linear variability, and frequently have fractal properties. When the fractal dimension of a time series is non‐integer, this is associated with two features: (1) inhomogeneity—extreme fluctuations at irregular intervals, and (2) scaling symmetries—proportionality relationships between fluctuations over different separation distances. In multivariate systems such as financial markets, fractality is stochastic rather than deterministic, and generally originates as a result of multiplicative interactions. Volatility diffusion models with multiple stochastic factors can generate fractal structures. In some cases, such as exchange rates, the underlying structural equation also gives rise to fractality. Fractal principles can be used to develop forecasting algorithms. The forecasting method that yields the best results here is the state transition‐fitted residual scale ratio (ST‐FRSR) model. A state transition model is used to predict the conditional probability of extreme events. Ratios of rates of change at proximate separation distances are used to parameterize the scaling symmetries. Forecasting experiments are run using intraday exchange rate futures contracts measured at 15‐minute intervals. The overall forecast error is reduced on average by up to 7% and in one instance by nearly a quarter. However, the forecast error during the outlying events is reduced by 39% to 57%. The ST‐FRSR reduces the predictive error primarily by capturing extreme fluctuations more accurately. Copyright © 2004 John Wiley & Sons, Ltd.     

Selection of the relevant information set for predictive relationships analysis between time series

In time series analysis, a vector Y is often called causal for another vector X if the former helps to improve the k‐step‐ahead forecast of the latter. If this holds for k=1, vector Y is commonly called Granger‐causal for X . It has been shown in several studies that the finding of causality between two (vectors of) variables is not robust to changes of the information set. In this paper, using the concept of Hilbert spaces, we derive a condition under which the predictive relationships between two vectors are invariant to the selection of a bivariate or trivariate framework. In more detail, we provide a condition under which the finding of causality (improved predictability at forecast horizon 1) respectively non‐causality of Y for X is unaffected if the information set is either enlarged or reduced by the information in a third vector Z . This result has a practical usefulness since it provides a guidance to validate the choice of the bivariate system { X , Y } in place of { X , Y , Z }. In fact, to test the ‘goodness’ of { X , Y } we should test whether Z Granger cause X not requiring the joint analysis of all variables in { X , Y , Z }. Copyright © 2002 John Wiley & Sons, Ltd.     

The versatility of spectrum analysis for forecasting financial time series

The versatility of the one‐dimensional discrete wavelet analysis combined with wavelet and Burg extensions for forecasting financial times series with distinctive properties is illustrated with market data. Any time series of financial assets may be decomposed into simpler signals called approximations and details in the framework of the one‐dimensional discrete wavelet analysis. The simplified signals are recomposed after extension. The final output is the forecasted time series which is compared to observed data. Results show the pertinence of adding spectrum analysis to the battery of tools used by econometricians and quantitative analysts for the forecast of economic or financial time series.     

Functional methods for time series prediction: a nonparametric approach

The problem of prediction in time series using nonparametric functional techniques is considered. An extension of the local linear method to regression with functional explanatory variable is proposed. This forecasting method is compared with the functional Nadaraya–Watson method and with finite‐dimensional nonparametric predictors for several real‐time series. Prediction intervals based on the bootstrap and conditional distribution estimation for those nonparametric methods are also compared. Copyright © 2010 John Wiley & Sons, Ltd.     

Bayesian analysis of concurrent time series with application to regional IBM revenue data

Business data frequently arise in the form of concurrent time series. We present a general framework for simultaneous modeling and fitting of such series using the class of Box—Jenkins models. This framework is an exchangeable hierarchical Bayesian model incorporating dependence among the series. Our motivating data set consists of regional IBM revenue available monthly for several geographic regions. Stationary seasonal autoregressive models are simultaneously fit to the regional data series using various error covariance specifications for the strong interregional dependence. A modified Gibbs sampling algorithm is used to carry out the fitting and to enable all subsequent inference. Graphical techniques using predictive distributions are employed to assess model adequacy and to select among models. Outlier estimation and prediction under the chosen model are used for planning and to measure the effect of special promotional events.     

Forecasting Australia's real house price index: A comparison of time series and machine learning methods

We employ 47 different algorithms to forecast Australian log real house prices and growth rates, and compare their ability to produce accurate out-of-sample predictions. The algorithms, which are specified in both single- and multi-equation frameworks, consist of traditional time series models, machine learning (ML) procedures, and deep learning neural networks. A method is adopted to compute iterated multistep forecasts from nonlinear ML specifications. While the rankings of forecast accuracy depend on the length of the forecast horizon, as well as on the choice of the dependent variable (log price or growth rate), a few generalizations can be made. For one- and two-quarter-ahead forecasts we find a large number of algorithms that outperform the random walk with drift benchmark. We also report several such outperformances at longer horizons of four and eight quarters, although these are not statistically significant at any conventional level. Six of the eight top forecasts (4 horizons × 2 dependent variables) are generated by the same algorithm, namely a linear support vector regressor (SVR). The other two highest ranked forecasts are produced as simple mean forecast combinations. Linear autoregressive moving average and vector autoregression models produce accurate olne-quarter-ahead predictions, while forecasts generated by deep learning nets rank well across medium and long forecast horizons.