|
on Forecasting |
| By: | Sebastian J. Dietz; Philipp Kneringer; Georg J. Mayr; Achim Zeileis |
| Abstract: | Low visibility conditions enforce special procedures that reduce the operational flight capacity at airports. Accurate and probabilistic forecasts of these capacity-reducing low-visibility procedure (lvp) states help the air traffic management to optimize flight planning and regulation. In this paper we investigate nowcasts, medium-range forecasts, and the predictability limit of the lvp states at Vienna Airport. The forecasts are computed with boosting trees, which consist of an ensemble of decision trees grown iteratively on residuals of previous trees. The model predictors are observations at Vienna Airport and output of a high resolution and an ensemble numerical weather prediction (NWP) model. Observations have highest impact for nowcasts up to a lead time of two hours. Afterwards a mix of observations and NWP forecast variables generates the most accurate predictions. With lead times longer than eight hours NWP output dominates until the predictability limit is reached at +12 days. For lead times longer than two days ensemble output generates higher improvement than a single higher resolution. The most important predictors for lead times up to +18 hours are observations of lvp and dew point depression, as well as NWP dew point depression. At longer lead times dew point depression and evaporation from the NWP models are most important. |
| Keywords: | aviation meteorology, statistical forecasting, visibility, ceiling, boosting, decision tree |
| Date: | 2018–07 |
| URL: | http://d.repec.org/n?u=RePEc:inn:wpaper:2018-11&r=for |
| By: | Thorsten Simon; Georg J. Mayr; Nikolaus Umlauf; Achim Zeileis |
| Abstract: | A method to predict lightning by postprocessing numerical weather prediction (NWP) output is developed for the region of the European Eastern Alps. Cloud-to-ground flashes-detected by the ground-based ALDIS network-are counted on the 18x18 km^2 grid of the 51-member NWP ensemble of the European Centre of Medium-Range Weather Forecasts (ECMWF). These counts serve as target quantity in count data regression models for the occurrence and the intensity of lightning events. The probability whether lightning occurs or not is modelled by a binomial distribution. For the intensity a hurdle approach is employed, for which the binomial distribution is combined with a zero-truncated negative binomial to model the counts within a grid cell. In both statistical models the parameters of the distributions are described by additive predictors, which are assembled by potentially nonlinear terms of NWP covariates. Measures of location and spread of approx. 100 direct and derived NWP covariates provide a pool of candidates for the nonlinear terms. A combination of stability selection and gradient boosting selects influential terms. Markov chain Monte Carlo (MCMC) simulation estimates the final model to provide credible inference of effects, scores and predictions. The selection of terms and MCMC simulation are applied for data of the year 2016, and out-of-sample performance is evaluated for 2017. The occurrence model outperforms a reference climatology-based on seven years of data-up to a forecast horizon of 5 days. The intensity model is calibrated and also outperforms climatology for exceedance probabilities, quantiles, and full predictive distributions. |
| Keywords: | lightning detection data, distributional regression, count data model, gradient boosting, MCMC |
| JEL: | C11 C53 Q54 |
| Date: | 2018–07 |
| URL: | http://d.repec.org/n?u=RePEc:inn:wpaper:2018-14&r=for |
| By: | Florian Huber; Gregor Kastner |
| Abstract: | We forecast S&P 500 excess returns using a flexible econometric state space model with non-Gaussian features at several levels. Estimation and prediction are conducted using fully-fledged Bayesian techniques. More precisely, we control for overparameterization via novel global-local shrinkage priors on the state innovation variances as well as the time-invariant part of the state space model. The shrinkage priors are complemented by heavy tailed state innovations that cater for potential large swings in the latent states even if the amount of shrinkage introduced is high. Moreover, we allow for leptokurtic stochastic volatility in the observation equation. The empirical findings indicate that several variants of the proposed approach outperform typical competitors frequently used in the literature, both in terms of point and density forecasts. Furthermore, a simple trading exercise shows that our framework also fares well when used for investment decisions. |
| Date: | 2018–05 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:1805.12217&r=for |
| By: | Alexey Ivashchenko (University of Lausanne and Swiss Finance Institute) |
| Abstract: | The part of credit spread that is not explained by corporate credit risk forecasts future economic activity. I show that the link with aggregate business risk and bond liquidity risk explains this fi nding. Once I project spreads on these two risk factors, which are readily measurable with the daily frequency, in addition to corporate credit risk, the forecasting power of the residual spread reduces substantially for some macro variables and disappears entirely for the others. Such residual, however, turns out to be an out-of-sample forecast of corporate bond market returns. An investment strategy based on such forecasts delivers risk-adjusted returns 50% higher than the corporate bond market. |
| Keywords: | credit spreads, corporate bond returns, business cycle, predictability of returns |
| JEL: | E44 G12 G17 |
| Date: | 2017–04 |
| URL: | http://d.repec.org/n?u=RePEc:chf:rpseri:rp1767&r=for |
| By: | Dominique Guégan (Université Paris1 Panthéon-Sorbonne, Centre d'Economie de la Sorbonne, LabEx ReFi and Ca' Foscari University of Venezia); Matteo Iacopini (Université Paris1 Panthéon-Sorbonne, Centre d'Economie de la Sorbonne, LabEx ReFi and Ca' Foscari University of Venezia) |
| Abstract: | The study of dependence between random variables is the core of theoretical and applied statistics. Static and dynamic copula models are useful for describing the dependence structure, which is fully encrypted in the copula probability density function. However, these models are not always able to describe the temporal change of the dependence patterns, which is a key characteristic of financial data. We propose a novel nonparametric framework for modelling a time series of copula probability density functions, which allows to forecast the entire function without the need of post-processing procedures to grant positiveness and unit integral. We exploit a suitable isometry that allows to transfer the analysis in a subset of the space of square integrable functions, where we build on nonparametric functional data analysis techniques to perform the analysis. The framework does not assume the densities to belong to any parametric family and it can be successfully applied also to general multivariate probability density functions with bounded or unbounded support. Finally, a noteworthy field of application pertains the study of time varying networks represented through vine copula models. We apply the proposed methodology for estimating and forecasting the time varying dependence structure between the S&P500 and NASDAQ indices |
| Keywords: | multivariate densities; functional PCA; nonparametric statistics; copula; functional time series; forecast; unbounded support |
| JEL: | C14 C53 |
| Date: | 2018–03 |
| URL: | http://d.repec.org/n?u=RePEc:mse:cesdoc:18012&r=for |
| By: | Gálvez-Soriano Oscar de Jesús |
| Abstract: | This paper evaluates five Nowcasting models that forecast Mexico's quarterly GDP: a Dynamic Factor Model (MFD), two Bridge Equation Models (BE) and two Principal Components Models (PCA). The results indicate that the average of the BE forecasts is statistically better than the rest of the models under consideration, according to the Diebold-Mariano (1995) accuracy test. In addition, using real-time information, the BE average is found to be more accurate than the median of the forecasts provided by the analysts surveyed by Bloomberg and the median of the experts who answer Banco de México's Survey of Professional Forecasters. |
| Keywords: | Nowcasting;Dynamic Factor Model;Bridge Equations;Principal Component Analysis;Quarterly GDP;Diebold-Mariano test |
| JEL: | C32 C38 C53 E52 |
| Date: | 2018–06 |
| URL: | http://d.repec.org/n?u=RePEc:bdm:wpaper:2018-06&r=for |
| By: | Kotilainen, Markku |
| Abstract: | The Research Institute of the Finnish Economy (Etla) started its business cycle forecasting activity in 1971. Forecasts have thus been published for almost half of a century. The foundations of forecasting have remained the same during this time. Some changes have, however, been implemented in the organization of the activity and in the tools used. In this memorandum the development of forecasting activity is described from its beginning. About twenty years the forecasts were produced by a matrix organization. The institute was divided into thematic research groups of which each produced the forecasts of its sector during the forecasting process. In 1989 a separate forecasting group was founded. At first, it was a part of a research program called ”forecasting activity”. Later, it was organized as a part of the research program ”macroeconomy, international economy and business cycles”. In the middle of the forecasting process have been the calculation framework based on the national accounts and the macroeconomic model of the institute. At first, these were integrated to each other. Nowadays they are separate. During the second half of the 1990s, the international macroeconomic model NiGEM was taken into use in forecasting and simulation of the international economy. Because Etla’s forecasts are detailed in terms of branches of industry, the institute’s input-output model is an important tool in producing the output forecasts. The main forecast publication is Suhdanne that has been published 2 times and occasionally even 4 times a year. In the 1990s the whole book was published in English, too, later just the extended summary. Since September 2016 Suhdanne is published also as an internet version, in addition to the paper one. |
| Keywords: | Research Institute of the Finnish Economy (Etla), forecasting, forecasting models |
| JEL: | E0 E17 E6 |
| Date: | 2018–07–11 |
| URL: | http://d.repec.org/n?u=RePEc:rif:briefs:69&r=for |
| By: | Yu-Long Zhou; Ren-Jie Han; Qian Xu; Wei-Ke Zhang |
| Abstract: | Intense volatility in financial markets affect humans worldwide. Therefore, relatively accurate prediction of volatility is critical. We suggest that massive data sources resulting from human interaction with the Internet may offer a new perspective on the behavior of market participants in periods of large market movements. First we select 28 key words, which are related to finance as indicators of the public mood and macroeconomic factors. Then those 28 words of the daily search volume based on Baidu index are collected manually, from June 1, 2006 to October 29, 2017. We apply a Long Short-Term Memory neural network to forecast CSI300 volatility using those search volume data. Compared to the benchmark GARCH model, our forecast is more accurate, which demonstrates the effectiveness of the LSTM neural network in volatility forecasting. |
| Date: | 2018–05 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:1805.11954&r=for |
| By: | Dominique Guegan (UP1 - Université Panthéon-Sorbonne, CES - Centre d'économie de la Sorbonne - UP1 - Université Panthéon-Sorbonne - CNRS - Centre National de la Recherche Scientifique, Labex ReFi - UP1 - Université Panthéon-Sorbonne, University of Ca’ Foscari [Venice, Italy]); Matteo Iacopini (UP1 - Université Panthéon-Sorbonne, CES - Centre d'économie de la Sorbonne - UP1 - Université Panthéon-Sorbonne - CNRS - Centre National de la Recherche Scientifique, Labex ReFi - UP1 - Université Panthéon-Sorbonne, University of Ca’ Foscari [Venice, Italy]) |
| Abstract: | The study of dependence between random variables is the core of theoretical and applied statistics. Static and dynamic copula models are useful for describing the dependence structure, which is fully encrypted in the copula probability density function. However, these models are not always able to describe the temporal change of the dependence patterns, which is a key characteristic of financial data. We propose a novel nonparametric framework for modelling a time series of copula probability density functions, which allows to forecast the entire function without the need of post-processing procedures to grant positiveness and unit integral. We exploit a suitable isometry that allows to transfer the analysis in a subset of the space of square integrable functions, where we build on nonparametric functional data analysis techniques to perform the analysis. The framework does not assume the densities to belong to any parametric family and it can be successfully applied also to general multivariate probability density functions with bounded or unbounded support. Finally, a noteworthy field of application pertains the study of time varying networks represented through vine copula models. We apply the proposed methodology for estimating and forecasting the time varying dependence structure between the S&P500 and NASDAQ indices. |
| Keywords: | nonparametric statistics,functional PCA,multivariate densities,copula,functional time series,forecast,unbounded support |
| Date: | 2018–03 |
| URL: | http://d.repec.org/n?u=RePEc:hal:cesptp:halshs-01821815&r=for |