Non‐linear,non‐parametric,non‐fundamental exchange rate forecasting

This paper employs a non‐parametric method to forecast high‐frequency Canadian/US dollar exchange rate. The introduction of a microstructure variable, order flow, substantially improves the predictive power of both linear and non‐linear models. The non‐linear models outperform random walk and linear models based on a number of recursive out‐of‐sample forecasts. Two main criteria that are applied to evaluate model performance are root mean squared error (RMSE) and the ability to predict the direction of exchange rate moves. The artificial neural network (ANN) model is consistently better in RMSE to random walk and linear models for the various out‐of‐sample set sizes. Moreover, ANN performs better than other models in terms of percentage of correctly predicted exchange rate changes. The empirical results suggest that optimal ANN architecture is superior to random walk and any linear competing model for high‐frequency exchange rate forecasting. Copyright © 2006 John Wiley & Sons, Ltd.     

A Neuro‐wavelet Model for the Short‐Term Forecasting of High‐Frequency Time Series of Stock Returns

We propose a wavelet neural network (neuro‐wavelet) model for the short‐term forecast of stock returns from high‐frequency financial data. The proposed hybrid model combines the capability of wavelets and neural networks to capture non‐stationary nonlinear attributes embedded in financial time series. A comparison study was performed on the predictive power of two econometric models and four recurrent neural network topologies. Several statistical measures were applied to the predictions and standard errors to evaluate the performance of all models. A Jordan net that used as input the coefficients resulting from a non‐decimated wavelet‐based multi‐resolution decomposition of an exogenous signal showed a consistent superior forecasting performance. Reasonable forecasting accuracy for the one‐, three‐ and five step‐ahead horizons was achieved by the proposed model. The procedure used to build the neuro‐wavelet model is reusable and can be applied to any high‐frequency financial series to specify the model characteristics associated with that particular series. Copyright © 2013 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.     

Adaptive Evolutionary Neural Networks for Forecasting and Trading without a Data‐Snooping Bias

In this paper, we present two neural‐network‐based techniques: an adaptive evolutionary multilayer perceptron (aDEMLP) and an adaptive evolutionary wavelet neural network (aDEWNN). The two models are applied to the task of forecasting and trading the SPDR Dow Jones Industrial Average (DIA), the iShares NYSE Composite Index Fund (NYC) and the SPDR S&P 500 (SPY) exchange‐traded funds (ETFs). We benchmark their performance against two traditional MLP and WNN architectures, a smooth transition autoregressive model (STAR), a moving average convergence/divergence model (MACD) and a random walk model. We show that the proposed architectures present superior forecasting and trading performance compared to the benchmarks and are free from the limitations of the traditional neural networks such as the data‐snooping bias and the time‐consuming and biased processes involved in optimizing their parameters. Copyright © 2015 John Wiley & Sons, Ltd.     

Why do regime‐switching models forecast so badly?

Most non‐linear techniques give good in‐sample fits to exchange rate data but are usually outperformed by random walks or random walks with drift when used for out‐of‐sample forecasting. In the case of regime‐switching models it is possible to understand why forecasts based on the true model can have higher mean squared error than those of a random walk or random walk with drift. In this paper we provide some analytical results for the case of a simple switching model, the segmented trend model. It requires only a small misclassification, when forecasting which regime the world will be in, to lose any advantage from knowing the correct model specification. To illustrate this we discuss some results for the DM/dollar exchange rate. We conjecture that the forecasting result is more general and describes limitations to the use of switching models for forecasting. This result has two implications. First, it questions the leading role of the random walk hypothesis for the spot exchange rate. Second, it suggests that the mean square error is not an appropriate way to evaluate forecast performance for non‐linear models. Copyright © 1999 John Wiley & Sons, Ltd.     

Building neural network models for time series: a statistical approach

This paper is concerned with modelling time series by single hidden layer feedforward neural network models. A coherent modelling strategy based on statistical inference is presented. Variable selection is carried out using simple existing techniques. The problem of selecting the number of hidden units is solved by sequentially applying Lagrange multiplier type tests, with the aim of avoiding the estimation of unidentified models. Misspecification tests are derived for evaluating an estimated neural network model. All the tests are entirely based on auxiliary regressions and are easily implemented. A small‐sample simulation experiment is carried out to show how the proposed modelling strategy works and how the misspecification tests behave in small samples. Two applications to real time series, one univariate and the other multivariate, are considered as well. Sets of one‐step‐ahead forecasts are constructed and forecast accuracy is compared with that of other nonlinear models applied to the same series. Copyright © 2006 John Wiley & Sons, Ltd.     

Long‐term streamflow forecasting using artificial neural network based on preprocessing technique

Artificial neural network (ANN) combined with signal decomposing methods is effective for long‐term streamflow time series forecasting. ANN is a kind of machine learning method utilized widely for streamflow time series, and which performs well in forecasting nonstationary time series without the need of physical analysis for complex and dynamic hydrological processes. Most studies take multiple factors determining the streamflow as inputs such as rainfall. In this study, a long‐term streamflow forecasting model depending only on the historical streamflow data is proposed. Various preprocessing techniques, including empirical mode decomposition (EMD), ensemble empirical mode decomposition (EEMD) and discrete wavelet transform (DWT), are first used to decompose the streamflow time series into simple components with different timescale characteristics, and the relation between these components and the original streamflow at the next time step is analyzed by ANN. Hybrid models EMD‐ANN, EEMD‐ANN and DWT‐ANN are developed in this study for long‐term daily streamflow forecasting, and performance measures root mean square error (RMSE), mean absolute percentage error (MAPE) and Nash–Sutcliffe efficiency (NSE) indicate that the proposed EEMD‐ANN method performs better than EMD‐ANN and DWT‐ANN models, especially in high flow forecasting.     

Forecasting Performance of Nonlinear Models for Intraday Stock Returns

We studied the predictability of intraday stock market returns using both linear and nonlinear time series models. For the S&P 500 index we compared simple autoregressive and random walk linear models with a range of nonlinear models, including smooth transition, Markov switching, artificial neural network, nonparametric kernel regression and support vector machine models for horizons of 5, 10, 20, 30 and 60 minutes. The empirical results indicate that nonlinear models outperformed linear models on the basis of both statistical and economic criteria. Specifically, although return serial correlation receded by around 10 minutes, return predictability still persisted for up to 60 minutes according to nonlinear models, even though profitability decreases as time elapses. More flexible nonlinear models such as support vector machines and artificial neural network did not clearly outperform other nonlinear models. Copyright © 2011 John Wiley & Sons, Ltd.     

Performance evaluation of neural network architectures: the case of predicting foreign exchange correlations

In the last decade, neural networks have emerged from an esoteric instrument in academic research to a rather common tool assisting auditors, investors, portfolio managers and investment advisors in making critical financial decisions. It is apparent that a better understanding of the network's performance and limitations would help both researchers and practitioners in analysing real‐world problems. Unlike many existing studies which focus on a single type of network architecture, this study evaluates and compares the performance of models based on two competing neural network architectures, the multi‐layered feedforward neural network (MLFN) and general regression neural network (GRNN). Our empirical evaluation measures the network models' strength on the prediction of currency exchange correlation with respect to a variety of statistical tests including RMSE, MAE, U statistic, Theil's decomposition test, Henriksson–Merton market timing test and Fair–Shiller informational content test. Results of experiments suggest that the selection of proper architectural design may contribute directly to the success in neural network forecasting. In addition, market timing tests indicate that both MLFN and GRNN models have economically significant values in predicting the exchange rate correlation. On the other hand, informational content tests discover that the neural network models based on different architectures capture useful information not found in each other and the information sets captured by the two network designs are independent of one another. An auxiliary experiment is developed and confirms the possible synergetic effect from combining forecasts made by the two different network architectures and from incorporating information from an implied correlation model into the neural network forecasts. Implied correlation and random walk models are also included in our empirical experiment for benchmark comparison. Copyright © 2005 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.     

Modelling and Trading the English and German Stock Markets with Novelty Optimization Techniques

The motivation for this paper was the introduction of novel short‐term models to trade the FTSE 100 and DAX 30 exchange‐traded funds (ETF) indices. There are major contributions in this paper which include the introduction of an input selection criterion when utilizing an expansive universe of inputs, a hybrid combination of partial swarm optimizer (PSO) with radial basis function (RBF) neural networks, the application of a PSO algorithm to a traditional autoregressive moving model (ARMA), the application of a PSO algorithm to a higher‐order neural network and, finally, the introduction of a multi‐objective algorithm to optimize statistical and trading performance when trading an index. All the machine learning‐based methodologies and the conventional models are adapted and optimized to model the index. A PSO algorithm is used to optimize the weights in a traditional RBF neural network, in a higher‐order neural network (HONN) and the AR and MA terms of an ARMA model. In terms of checking the statistical and empirical accuracy of the novel models, we benchmark them with a traditional HONN, with an ARMA, with a moving average convergence/divergence model (MACD) and with a naïve strategy. More specifically, the trading and statistical performance of all models is investigated in a forecast simulation of the FTSE 100 and DAX 30 ETF time series over the period January 2004 to December 2015 using the last 3 years for out‐of‐sample testing. Finally, the empirical and statistical results indicate that the PSO‐RBF model outperforms all other examined models in terms of trading accuracy and profitability, even with mixed inputs and with only autoregressive inputs. Copyright © 2016 John Wiley & Sons, Ltd.     

Conditional predictability of daily exchange rates

At what forecast horizon is one time series more predictable than another? This paper applies the Diebold–Kilian conditional predictability measure to assess the out‐of‐sample performance of three alternative models of daily GBP/USD and DEM/USD exchange rate returns. Predictability is defined as a non‐linear statistic of a model's relative expected losses at short and long forecast horizons, allowing flexible choice of both the estimation procedure and loss function. The long horizon is set to 2 weeks and one month ahead and forecasts evaluated according to MSE loss. Bootstrap methodology is used to estimate the data's conditional predictability using GARCH models. This is then compared to predictability under a random walk and a model using the prediction bias in uncovered interest parity (UIP). We find that both exchange rates are less predictable using GARCH than using a random walk, but they are more predictable using UIP than a random walk. Predictability using GARCH is relatively higher for the 2‐weeks‐than for the 1‐month long forecast horizon. Comparing the results using a random walk to that using UIP reveals ‘pockets’ of predictability, that is, particular short horizons for which predictability using the random walk exceeds that using UIP, or vice versa. Overall, GBP/USD returns appear more predictable than DEM/USD returns at short horizons. Copyright © 2002 John Wiley & Sons, Ltd.     

Foreign exchange market prediction with multiple classifiers

Foreign exchange market prediction is attractive and challenging. According to the efficient market and random walk hypotheses, market prices should follow a random walk pattern and thus should not be predictable with more than about 50% accuracy. In this article, we investigate the predictability of foreign exchange spot rates of the US dollar against the British pound to show that not all periods are equally random. We used the Hurst exponent to select a period with great predictability. Parameters for generating training patterns were determined heuristically by auto‐mutual information and false nearest‐neighbor methods. Some inductive machine‐learning classifiers—artificial neural network, decision tree, k‐nearest neighbor, and naïve Bayesian classifier—were then trained with these generated patterns. Through appropriate collaboration of these models, we achieved a prediction accuracy of up to 67%. Copyright © 2009 John Wiley & Sons, Ltd.     

Ensemble Forecasting for Complex Time Series Using Sparse Representation and Neural Networks

Based on the concept of ‘decomposition and ensemble’, a novel ensemble forecasting approach is proposed for complex time series by coupling sparse representation (SR) and feedforward neural network (FNN), i.e. the SR‐based FNN approach. Three main steps are involved: data decomposition via SR, individual forecasting via FNN and ensemble forecasting via a simple addition method. In particular, to capture various coexisting hidden factors, the effective decomposition tool of SR with its unique virtues of flexibility and generalization is introduced to formulate an overcomplete dictionary covering diverse bases, e.g. exponential basis for main trend, Fourier basis for cyclical (and seasonal) features and wavelet basis for transient actions, different from other techniques with a single basis. Using crude oil price (a typical complex time series) as sample data, the empirical study statistically confirms the superiority of the SR‐based FNN method over some other popular forecasting models and similar ensemble models (with other decomposition tools). Copyright © 2016 John Wiley & Sons, Ltd.     

Non‐linear forecasts of stock returns

Following recent non‐linear extensions of the present‐value model, this paper examines the out‐of‐sample forecast performance of two parametric and two non‐parametric nonlinear models of stock returns. The parametric models include the standard regime switching and the Markov regime switching, whereas the non‐parametric are the nearest‐neighbour and the artificial neural network models. We focused on the US stock market using annual observations spanning the period 1872–1999. Evaluation of forecasts was based on two criteria, namely forecast accuracy and forecast encompassing. In terms of accuracy, the Markov and the artificial neural network models produce at least as accurate forecasts as the other models. In terms of encompassing, the Markov model outperforms all the others. Overall, both criteria suggest that the Markov regime switching model is the most preferable non‐linear empirical extension of the present‐value model for out‐of‐sample stock return forecasting. Copyright © 2003 John Wiley & Sons, Ltd.     

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.     

The performance of non‐linear exchange rate models: a forecasting comparison

In recent years there has been a considerable development in modelling non‐linearities and asymmetries in economic and financial variables. The aim of the current paper is to compare the forecasting performance of different models for the returns of three of the most traded exchange rates in terms of the US dollar, namely the French franc (FF/$), the German mark (DM/$) and the Japanese yen (Y/$). The relative performance of non‐linear models of the SETAR, STAR and GARCH types is contrasted with their linear counterparts. The results show that if attention is restricted to mean square forecast errors, the performance of the models, when distinguishable, tends to favour the linear models. The forecast performance of the models is evaluated also conditional on the regime at the forecast origin and on density forecasts. This analysis produces more evidence of forecasting gains from non‐linear models. Copyright © 2002 John Wiley & Sons, Ltd.     

A Robust Data‐Mining Approach to Bankruptcy Prediction

In this study, new variants of genetic programming (GP), namely gene expression programming (GEP) and multi‐expression programming (MEP), are utilized to build models for bankruptcy prediction. Generalized relationships are obtained to classify samples of 136 bankrupt and non‐bankrupt Iranian corporations based on their financial ratios. An important contribution of this paper is to identify the effective predictive financial ratios on the basis of an extensive bankruptcy prediction literature review and upon a sequential feature selection analysis. The predictive performance of the GEP and MEP forecasting methods is compared with the performance of traditional statistical methods and a generalized regression neural network. The proposed GEP and MEP models are effectively capable of classifying bankrupt and non‐bankrupt firms and outperform the models developed using other methods. Copyright © 2011 John Wiley & Sons, Ltd.     

Forecast Evaluation of Nonlinear Models: The Case of Long‐Span Real Exchange Rates

This paper deals with the nonlinear modeling and forecasting of the dollar–sterling and franc–sterling real exchange rates using long spans of data. Our contribution is threefold. First, we provide significant evidence of smooth transition dynamics in the series by employing a battery of recently developed in‐sample statistical tests. Second, we investigate the small‐sample properties of several evaluation measures for comparing recursive forecasts when one of the competing models is nonlinear. Finally, we run a forecasting race for the post‐Bretton Woods era between the nonlinear real exchange rate model, the random walk, and the linear autoregressive model. The nonlinear model outperforms all rival models in the dollar–sterling case but cannot beat the linear autoregressive in the franc–sterling. Copyright © 2011 John Wiley & Sons, Ltd.     

Covariance estimation for multivariate conditionally Gaussian dynamic linear models

In multivariate time series, estimation of the covariance matrix of observation innovations plays an important role in forecasting as it enables computation of standardized forecast error vectors as well as the computation of confidence bounds of forecasts. We develop an online, non‐iterative Bayesian algorithm for estimation and forecasting. It is empirically found that, for a range of simulated time series, the proposed covariance estimator has good performance converging to the true values of the unknown observation covariance matrix. Over a simulated time series, the new method approximates the correct estimates, produced by a non‐sequential Monte Carlo simulation procedure, which is used here as the gold standard. The special, but important, vector autoregressive (VAR) and time‐varying VAR models are illustrated by considering London metal exchange data consisting of spot prices of aluminium, copper, lead and zinc. Copyright © 2007 John Wiley & Sons, Ltd.