Forecasting using locally stationary wavelet processes

Locally stationary wavelet (LSW) processes, built on non-decimated wavelets, can be used to analyse and forecast non-stationary time series. They have been proved useful in the analysis of financial data. In this paper, we first carry out a sensitivity analysis, then propose some practical guidelines for choosing the wavelet bases for these processes. The existing forecasting algorithm is found to be vulnerable to outliers, and a new algorithm is proposed to overcome the weakness. The new algorithm is shown to be stable and outperforms the existing algorithm when applied to real financial data. The volatility forecasting ability of LSW modelling based on our new algorithm is then discussed and shown to be competitive with traditional GARCH models.     

Functional time series approach for forecasting very short-term electricity demand

This empirical paper presents a number of functional modelling and forecasting methods for predicting very short-term (such as minute-by-minute) electricity demand. The proposed functional methods slice a seasonal univariate time series (TS) into a TS of curves; reduce the dimensionality of curves by applying functional principal component analysis before using a univariate TS forecasting method and regression techniques. As data points in the daily electricity demand are sequentially observed, a forecast updating method can greatly improve the accuracy of point forecasts. Moreover, we present a non-parametric bootstrap approach to construct and update prediction intervals, and compare the point and interval forecast accuracy with some naive benchmark methods. The proposed methods are illustrated by the half-hourly electricity demand from Monday to Sunday in South Australia.     

A robust functional time series forecasting method

Univariate time series often take the form of a collection of curves observed sequentially over time. Examples of these include hourly ground-level ozone concentration curves. These curves can be viewed as a time series of functions observed at equally spaced intervals over a dense grid. Since functional time series may contain various types of outliers, we introduce a robust functional time series forecasting method to down-weigh the influence of outliers in forecasting. Through a robust principal component analysis based on projection pursuit, a time series of functions can be decomposed into a set of robust dynamic functional principal components and their associated scores. Conditioning on the estimated functional principal components, the crux of the curve-forecasting problem lies in modelling and forecasting principal component scores, through a robust vector autoregressive forecasting method. Via a simulation study and an empirical study on forecasting ground-level ozone concentration, the robust method demonstrates the superior forecast accuracy that dynamic functional principal component regression entails. The robust method also shows the superior estimation accuracy of the parameters in the vector autoregressive models for modelling and forecasting principal component scores, and thus improves curve forecast accuracy.     

Lee–Carter mortality forecasting: a parallel generalized linear modelling approach for England and Wales mortality projections

Summary. The paper presents a reinterpretation of the model underpinning the Lee–Carter methodology for forecasting mortality (and other vital) rates. A parallel methodology based on generalized linear modelling is introduced. The use of residual plots is proposed for both methods to aid the assessment of the goodness of fit. The two methods are compared in terms of structure and assumptions. They are then compared through an analysis of the gender- and age-specific mortality rates for England and Wales over the period 1950–1998 and through a consideration of the forecasts generated by the two methods. The paper also compares different approaches to the forecasting of life expectancy and considers the effectiveness of the Coale–Guo method for extrapolating mortality rates to the oldest ages.     

Keep it simple: on specific-to-general predictor selection for time series forecasting in the short,medium and long run

This paper describes the forecasting performance of general-to-specific and specific-to-general predictor selection within specifications fitting into the class of (approximate) linear autoregressions. Short, medium and long horizon forecasting exercises are distinguished. Regarding the latter, iterative prediction is compared with direct conditioning on available time series information. Ex ante forecasting results are provided for 495 real macro-economic and financial time series recently collected for 25 economies and the Euro area [A. Inouea and L. Kilian, On the selection of forecasting models, J. Econ. 130 (2006), pp. 273–306]. Almost 9000 single predictions enter the modelling comparison. Overall, specific-to-general predictor selection turns out to offer preferable prediction outcomes in terms of statistical and more economic loss functions. With regard to medium (long) term prediction, the analysis is supportive for direct (iterative) multistep prediction.     

Nonparametric forecasting: a comparison of three kernel-based methods

In this paper the use of three kernel-based nonparametric forecasting methods - the conditional mean, the conditional median, and the conditional mode -is explored in detail. Several issues related to the estimation of these methods are discussed, including the choice of the bandwidth and the type of kernel function. The out-of-sample forecasting performance of the three nonparametric methods is investigated using 60 real time series. We find that there is no superior forecast method for series having approximately less than 100 observations. However, when a time series is long or when its conditional density is bimodal there is quite a difference between the forecasting performance of the three kernel-based forecasting methods.     

Evolving possibilistic fuzzy modelling

This paper suggests an evolving possibilistic approach for fuzzy modelling of time-varying processes. The approach is based on an extension of the well-known possibilistic fuzzy c-means (FCM) clustering and functional fuzzy rule-based modelling. Evolving possibilistic fuzzy modelling (ePFM) employs memberships and typicalities to recursively cluster data, and uses participatory learning to adapt the model structure as a stream data is input. The idea of possibilistic clustering plays a key role when the data are noisy and with outliers due to the relaxation of the restriction on membership degrees to add up unity in FCM clustering algorithm. To show the usefulness of ePFM, the approach is addressed for system identification using Box & Jenkins gas furnace data as well as time series forecasting considering the chaotic Mackey–Glass series and data produced by a synthetic time-varying process with parameter drift. The results show that ePFM is a potential candidate for nonlinear time-varying systems modelling, with comparable or better performance than alternative approaches, mainly when noise and outliers affect the data available.     

An Alternative Methodology for Combining Different Forecasting Models

Many economic and financial time series exhibit heteroskedasticity, where the variability changes are often based on recent past shocks, which cause large or small fluctuations to cluster together. Classical ways of modelling the changing variance include the use of Generalized Autoregressive Conditional Heteroskedasticity (GARCH) models and Neural Networks models. The paper starts with a comparative study of these two models, both in terms of capturing the non-linear or heteroskedastic structure and forecasting performance. Monthly and daily exchange rates for three different countries are implemented. The paper continues with different methods for combining forecasts of the volatility from the competing models, in order to improve forecasting accuracy. Traditional methods for combining the predicted values from different models, using various weighting schemes are considered, such as the simple average or methods that find the best weights in terms of minimizing the squared forecast error. The main purpose of the paper is, however, to propose an alternative methodology for combining forecasts effectively. The new, hereby-proposed non-linear, non-parametric, kernel-based method, is shown to have the basic advantage of not being affected by outliers, structural breaks or shocks to the system and it does not require a specific functional form for the combination.     

Estimation and forecasting hospital admissions due to Influenza: Planning for winter pressure. The case of the West Midlands, UK

Winters are a difficult period for the National Health Service (NHS) in the United Kingdom (UK), due to the combination of cold weather and the increased likelihood of respiratory infections, especially influenza. In this article we present a proper statistical time series approach for modelling and analysing weekly hospital admissions in the West Midlands in the UK during the period week 15/1990 to week 14/1999. We consider three variables, namely, hospital admissions, general practitioner consultants, and minimum temperature. The autocorrelations of each series are shown to decay hyperbolically. The correlations of hospital admission and the lag of other series also decay hyperbolically but with different speed and directions. One of the main objectives of this paper is to show that each of the three series can be represented by a Fractional Differenced Autoregressive integrated moving average model, (FDA). Further, the hospital admission winter and summer residuals shows significant interdependency, which may be interpreted as hidden periodicities within the last 10-years time interval. The short-range (8 weeks) forecasting of hospital admission of the FDA model and a fourth-order AutoRegressive AR(4) model are quite similar. However, our results reveal that the long-range forecasting of FDA is more realistic. This implies that, using the FDA approach, the respective authority can plan for winter pressure properly.     

Observable trend-projecting state-space models

Much attention has focused in recent years on the use of state-space models for describing and forecasting industrial time series. However, several state-space models that are proposed for such data series are not observable and do not have a unique representation, particularly in situations where the data history suggests marked seasonal trends. This raises major practical difficulties since it becomes necessary to impose one or more constraints and this implies a complicated error structure on the model. The purpose of this paper is to demonstrate that state-space models are useful for describing time series data for forecasting purposes and that there are trend-projecting state-space components that can be combined to provide observable state-space representations for specified data series. This result is particularly useful for seasonal or pseudo-seasonal time series. A well-known data series is examined in some detail and several observable state-space models are suggested and compared favourably with the constrained observable model.     

Observable trend-projecting state-space models

Much attention has focused in recent years on the use of state-space models for describing and forecasting industrial time series. However, several state-space models that are proposed for such data series are not observable and do not have a unique representation, particularly in situations where the data history suggests marked seasonal trends. This raises major practical difficulties since it becomes necessary to impose one or more constraints and this implies a complicated error structure on the model. The purpose of this paper is to demonstrate that state-space models are useful for describing time series data for forecasting purposes and that there are trend-projecting state-space components that can be combined to provide observable state-space representations for specified data series. This result is particularly useful for seasonal or pseudo-seasonal time series. A well-known data series is examined in some detail and several observable state-space models are suggested and compared favourably with the constrained observable model.     

Randomized singular spectrum analysis for long time series

ABSTRACT

Singular spectrum analysis (SSA) is a relatively new method for time series analysis and comes as a non-parametric alternative to the classical methods. This methodology has proven to be effective in analysing non-stationary and complex time series since it is a non-parametric method and do not require the classical assumptions over the stationarity or over the normality of the residuals. Although SSA have proved to provide advantages over traditional methods, the challenges that arise when long time series are considered, make the standard SSA very demanding computationally and often not suitable. In this paper we propose the randomized SSA which is an alternative to SSA for long time series without losing the quality of the analysis. The SSA and the randomized SSA are compared in terms of quality of the model fit and forecasting, and computational time. This is done by using Monte Carlo simulations and real data about the daily prices of five of the major world commodities.     

Short-term forecasting with a computationally efficient nonparametric transfer function model

In this paper a semi-parametric approach is developed to model non-linear relationships in time series data using polynomial splines. Polynomial splines require very little assumption about the functional form of the underlying relationship, so they are very flexible and can be used to model highly non-linear relationships. Polynomial splines are also computationally very efficient. The serial correlation in the data is accounted for by modelling the noise as an autoregressive integrated moving average (ARIMA) process, by doing so, the efficiency in nonparametric estimation is improved and correct inferences can be obtained. The explicit structure of the ARIMA model allows the correlation information to be used to improve forecasting performance. An algorithm is developed to automatically select and estimate the polynomial spline model and the ARIMA model through backfitting. This method is applied on a real-life data set to forecast hourly electricity usage. The non-linear effect of temperature on hourly electricity usage is allowed to be different at different hours of the day and days of the week. The forecasting performance of the developed method is evaluated in post-sample forecasting and compared with several well-accepted models. The results show the performance of the proposed model is comparable with a long short-term memory deep learning model.     

Efficient computation of the discrete autocorrelation wavelet inner product matrix

Discrete autocorrelation (a.c.) wavelets have recently been applied in the statistical analysis of locally stationary time series for local spectral modelling and estimation. This article proposes a fast recursive construction of the inner product matrix of discrete a.c. wavelets which is required by the statistical analysis. The recursion connects neighbouring elements on diagonals of the inner product matrix using a two-scale property of the a.c. wavelets. The recursive method is an (log (N)³) operation which compares favourably with the (N log N) operations required by the brute force approach. We conclude by describing an efficient construction of the inner product matrix in the (separable) two-dimensional case.     

INDEX TO VOLUME 8 (1 990)

This article provides a simple shrinkage representation that describes the operational characteristics of various forecasting methods designed for a large number of orthogonal predictors (such as principal components). These methods include pretest methods, Bayesian model averaging, empirical Bayes, and bagging. We compare empirically forecasts from these methods with dynamic factor model (DFM) forecasts using a U.S. macroeconomic dataset with 143 quarterly variables spanning 1960–2008. For most series, including measures of real economic activity, the shrinkage forecasts are inferior to the DFM forecasts. This article has online supplementary material.     

Coherent mortality forecasting by the weighted multilevel functional principal component approach

In human mortality modelling, if a population consists of several subpopulations it can be desirable to model their mortality rates simultaneously while taking into account the heterogeneity among them. The mortality forecasting methods tend to result in divergent forecasts for subpopulations when independence is assumed. However, under closely related social, economic and biological backgrounds, mortality patterns of these subpopulations are expected to be non-divergent in the future. In this article, we propose a new method for coherent modelling and forecasting of mortality rates for multiple subpopulations, in the sense of nondivergent life expectancy among subpopulations. The mortality rates of subpopulations are treated as multilevel functional data and a weighted multilevel functional principal component (wMFPCA) approach is proposed to model and forecast them. The proposed model is applied to sex-specific data for nine developed countries, and the results show that, in terms of overall forecasting accuracy, the model outperforms the independent model and the Product-Ratio model as well as the unweighted multilevel functional principal component approach.     

Forecasting Multivariate Road Traffic Flows Using Bayesian Dynamic Graphical Models,Splines and Other Traffic Variables

Traffic flow data are routinely collected for many networks worldwide. These invariably large data sets can be used as part of a traffic management system, for which good traffic flow forecasting models are crucial. The linear multiregression dynamic model (LMDM) has been shown to be promising for forecasting flows, accommodating multivariate flow time series, while being a computationally simple model to use. While statistical flow forecasting models usually base their forecasts on flow data alone, data for other traffic variables are also routinely collected. This paper shows how cubic splines can be used to incorporate extra variables into the LMDM in order to enhance flow forecasts. Cubic splines are also introduced into the LMDM to parsimoniously accommodate the daily cycle exhibited by traffic flows. The proposed methodology allows the LMDM to provide more accurate forecasts when forecasting flows in a real high‐dimensional traffic data set. The resulting extended LMDM can deal with some important traffic modelling issues not usually considered in flow forecasting models. Additionally, the model can be implemented in a real‐time environment, a crucial requirement for traffic management systems designed to support decisions and actions to alleviate congestion and keep traffic flowing.     

Boosting techniques for nonlinear time series models

Many of the popular nonlinear time series models require a priori the choice of parametric functions which are assumed to be appropriate in specific applications. This approach is mainly used in financial applications, when sufficient knowledge is available about the nonlinear structure between the covariates and the response. One principal strategy to investigate a broader class on nonlinear time series is the Nonlinear Additive AutoRegressive (NAAR) model. The NAAR model estimates the lags of a time series as flexible functions in order to detect non-monotone relationships between current and past observations. We consider linear and additive models for identifying nonlinear relationships. A componentwise boosting algorithm is applied for simultaneous model fitting, variable selection, and model choice. Thus, with the application of boosting for fitting potentially nonlinear models we address the major issues in time series modelling: lag selection and nonlinearity. By means of simulation we compare boosting to alternative nonparametric methods. Boosting shows a strong overall performance in terms of precise estimations of highly nonlinear lag functions. The forecasting potential of boosting is examined on the German industrial production (IP); to improve the model’s forecasting quality we include additional exogenous variables. Thus we address the second major aspect in this paper which concerns the issue of high dimensionality in models. Allowing additional inputs in the model extends the NAAR model to a broader class of models, namely the NAARX model. We show that boosting can cope with large models which have many covariates compared to the number of observations.     

Analyzing Permanent and Transient Influences in Multiple Time Series Models

Several multiple time series models are developed and applied to the analysis and forecasting of the M1 and M2 money supply aggregates. These models feature a decomposition of the time series into permanent and transient influences or components. This decomposition appears to enhance forecasting accuracy and is associated with a variance-covariance allocation parameter that is also estimated from the data. Conditional maximum likelihood estimates for model parameters are presented as well as a numerical algorithm that is an adaptation of Marquardt's algorithm.     

SEMIFAR forecasts, with applications to foreign exchange rates

SEMIFAR models introduced in Beran (1997. Estimating trends, long-range dependence and nonstationarity, preprint) provide a semiparametric modelling framework that enables the data analyst to separate deterministic and stochastic trends as well as short- and long-memory components in an observed time series. A correct distinction between these components, and in particular, the decision which of the components may be present in the data have an important impact on forecasts. In this paper, forecasts and forecast intervals for SEMIFAR models are obtained. The forecasts are based on an extrapolation of the nonparametric trend function and optimal forecasts of the stochastic component. In the data analytical part of the paper, the proposed method is applied to foreign exchange rates from Europe and Asia.