|
on Forecasting |
| By: | Shiyi Chen; Kiho Jeong; Wolfgang K. Härdle |
| Abstract: | Motivated by the recurrent Neural Networks, this paper proposes a recurrent Support Vector Regression (SVR) procedure to forecast nonlinear ARMA model based simulated data and real data of financial returns. The forecasting ability of the recurrent SVR is compared with three competing methods, MLE, recurrent MLP and feedforward SVR. Theoretically, MLE and MLP only focus on fit in-sample, but SVR considers both fit and forecast out-of-sample which endows SVR with an excellent forecasting ability. This is confirmed by the evidence from the simulated and real data based on two forecasting accuracy evaluation metrics (NSME and sign). That is, for one-step-ahead forecasting, the recurrent SVR is consistently better than the MLE and the recurrent MLP in forecasting both the magnitude and turning points, and really improves the forecasting performance as opposed to the usual feedforward SVR. |
| Keywords: | Recurrent Support Vector Regression; MLE; recurrent MLP; nonlinear ARMA; financial forecasting |
| JEL: | C45 F37 F47 |
| Date: | 2008–07 |
| URL: | http://d.repec.org/n?u=RePEc:hum:wpaper:sfb649dp2008-051&r=for |
| By: | Lillie Lam (Research Department, Hong Kong Monetary Authority); Laurence Fung (Research Department, Hong Kong Monetary Authority); Ip-wing Yu (Research Department, Hong Kong Monetary Authority) |
| Abstract: | Exchange-rate movement is regularly monitored by central banks for macroeconomic-analysis and market-surveillance purposes. Notwithstanding the pioneering study of Meese and Rogoff (1983), which shows the superiority of the random-walk model in out-of-sample exchange-rate forecast, there is some evidence that exchange-rate movement may be predictable at longer time horizons. This study compares the forecast performance of the Purchasing Power Parity model, Uncovered Interest Rate Parity model, Sticky Price Monetary model, the model based on the Bayesian Model Averaging technique, and a combined forecast of all the above models with benchmarks given by the random-walk model and the historical average return. Empirical results suggest that the combined forecast outperforms the benchmarks and generally yields better results than relying on a single model. |
| Keywords: | Bayesian Analysis, Model Evaluation and Selection, Forecasting and Other Model Application |
| JEL: | C11 C52 C53 |
| Date: | 2008–06 |
| URL: | http://d.repec.org/n?u=RePEc:hkg:wpaper:0808&r=for |
| By: | Riccardo Cristadoro (Bank of Italy, Economic Research Department); Fabrizio Venditti (Bank of Italy, Economic Research Department); Giuseppe Saporito (Bank of Italy, Cagliari) |
| Abstract: | The ECB objective of price stability is given a quantitative content as a year-on-year growth rate in the euro area HICP close but below 2% over the medium term. While this objective is referred to area-wide price developments, in anticipating monetary policy moves, market analysts pay considerable attention to national data. In this paper we use the Generalized Dynamic Factor Model to derive a set of core inflation indicators that, combining national with area-wide data, allow us to answer two related questions: whether country-specific data are actually relevant to the future path of area-wide inflation once the information contained in area-wide data has been exploited, and whether it is useful, in order to track ECB monetary policy decisions, to factor in national and not only area-wide statistics. In both cases, our findings suggest that, when area-wide information is properly taken into account, there is little to be gained by considering national idiosyncratic developments. |
| Keywords: | Forecast, Dynamic factor model, inflation, monetary policy |
| JEL: | C25 E37 E52 |
| Date: | 2008–06 |
| URL: | http://d.repec.org/n?u=RePEc:bdi:wptemi:td_677_08&r=for |
| By: | Gary Koop (University of Strathclyde, Glasgow, UK and The Rimini Centre for Economic Analysis, Italy); Markus Jochmann (University of Strathclyde, Glasgow, UK and The Rimini Centre for Economic Analysis, Italy); Rodney W. Strachan (University of Queensland, UK and The Rimini Centre for Economic Analysis, Italy) |
| Abstract: | This paper builds a model which has two extensions over a standard VAR. The first of these is stochastic search variable selection, which is an automatic model selection device which allows for coefficients in a possibly over-parameterized VAR to be set to zero. The second allows for an unknown number of structual breaks in the VAR parameters. We investigate the in-sample and forecasting performance of our model in an application involving a commonly-used US macro-economic data set. We find that, in-sample, these extensions clearly are warranted. In a recursive forecasting exercise, we find moderate improvements over a standard VAR, although most of these improvements are due to the use of stochastic search variable selection rather than the inclusion of breaks. Classification-JEL: |
| Date: | 2008–01 |
| URL: | http://d.repec.org/n?u=RePEc:rim:rimwps:19-08&r=for |
| By: | Dimitrios D. Thomakos (University of Peloponnese, Greece and The Rimini Centre for Economic Analysis) |
| Abstract: | In this paper I propose a novel optimal linear ølter for smoothing, trend and signal extraction for time series with a unit root. The filter is based on the Singular Spectrum Analysis (SSA) methodology, takes the form of a particular moving average and is di¨erent from other linear filters that have been used in the existing literature. To best of my knowledge this is the first time that moving average smoothing is given an optimality justification for use with unit root processes. The frequency response function of the filter is examined and a new method for selecting the degree of smoothing is suggested. I also show that the filter can be used for successfully extracting a unit root signal from stationary noise. The proposed methodology can be extended to also deal with two cointegrated series and I show how to estimate the cointegrating coe±cient using SSA and how to extract the common stochastic trend component. A simulation study explores some of the characteristics of the filter for signal extraction, trend prediction and cointegration estimation for univariate and bivariate series. The practical usefulness of the method is illustrated using data for the US real GDP and two financial time series. Classification-JEL: |
| Keywords: | cointegration, forecasting, linear øltering, singular spectrum analysis, smoothing, trend extraction and prediction, unit root. |
| Date: | 2008–01 |
| URL: | http://d.repec.org/n?u=RePEc:rim:rimwps:14-08&r=for |
| By: | Menzie D. Chinn; Michael J. Moore |
| Abstract: | We propose an exchange rate model which is a hybrid of the conventional specification with monetary fundamentals and the Evans-Lyons microstructure approach. It argues that the failure of the monetary model is principally due to private preference shocks which render the demand for money unstable. These shocks to liquidity preference are revealed through order flow. We estimate a model augmented with order flow variables, using a unique data set: almost 100 monthly observations on inter-dealer order flow on dollar/euro and dollar/yen. The augmented macroeconomic, or "hybrid", model exhibits out of sample forecasting improvement over the basic macroeconomic and random walk specifications. |
| JEL: | D82 F31 F41 F47 |
| Date: | 2008–07 |
| URL: | http://d.repec.org/n?u=RePEc:nbr:nberwo:14175&r=for |
| By: | Steven C. Bourassa (University of Louisville, School of Urban and Public Affairs); Eva Cantoni (University of Geneva, Departement of Econometrics); Martin Hoesli |
| Abstract: | We analyze the impacts of alternative submarket definitions when predicting house prices in a mass appraisal context, using both ordinary least squares (OLS) and geostatistical techniques. For this purpose, we use over 13,000 housing transactions for Louisville, Kentucky. We use districts defined by the local property tax assessment office as well as a classification of census tracts generated by principal components and cluster analysis. We also experiment with varying numbers of census tract groupings. Our results indicate that somewhat better results are obtained with more homogeneous submarkets. Also, the accuracy of house price predictions increases as the number of submarkets is increased, but then quickly levels off. Adding submarket variables to the OLS model yields price predictions that are as accurate as when geostatistical methods are used to account for spatial dependence in the error terms. However, using both dummy variables for submarkets and geostatistical methods leads to significant increases in accuracy. |
| Keywords: | spatial dependence, hedonic price models, geostatistical models, mass appraisal, housing submarkets. |
| JEL: | C11 D58 D84 D91 |
| Date: | 2008–01 |
| URL: | http://d.repec.org/n?u=RePEc:chf:rpseri:rp0801&r=for |