The Experts below are selected from a list of 33882 Experts worldwide ranked by ideXlab platform
Rafał Weron - One of the best experts on this subject based on the ideXlab platform.
-
On the importance of the long-term Seasonal Component in day-ahead electricity price forecasting. Part II – Probabilistic forecasting
Energy Economics, 2019Co-Authors: Bartosz Uniejewski, Grzegorz Marcjasz, Rafał WeronAbstract:A recent electricity price forecasting study has shown that the Seasonal Component AutoRegressive (SCAR) modeling framework, which consists of decomposing a series of spot prices into a trend-Seasonal and a stochastic Component, modeling them independently and then combining their forecasts, can yield more accurate point predictions than an approach in which the same autoregressive model is calibrated to the prices themselves. Here, we show that further accuracy gains can be achieved when the explanatory variables (load forecasts) are deSeasonalized as well. More importantly, considering a novel extension of the SCAR concept to probabilistic forecasting and applying two methods of combining predictive distributions, we find that (i) SCAR-type models nearly always significantly outperform the autoregressive benchmark but are in turn outperformed by combined SCAR forecasts, (ii) predictive distributions computed using Quantile Regression Averaging (QRA) outperform those obtained from historical simulation and bootstrap methods, and (iii) averaging over predictive distributions generally yields better probabilistic forecasts of electricity spot prices than averaging over quantiles. Given that probabilistic forecasting is a concept closely related to risk management, our study has important implications for risk officers and portfolio managers in the power sector.
-
On the importance of the long-term Seasonal Component in day-ahead electricity price forecasting with NARX neural networks
International Journal of Forecasting, 2019Co-Authors: Grzegorz Marcjasz, Bartosz Uniejewski, Rafał WeronAbstract:Abstract Daily and weekly Seasonalities are always taken into account in day-ahead electricity price forecasting, but the long-term Seasonal Component has long been believed to add unnecessary complexity, and hence, most studies have ignored it. The recent introduction of the Seasonal Component AutoRegressive (SCAR) modeling framework has changed this viewpoint. However, this framework is based on linear models estimated using ordinary least squares. This paper shows that considering non-linear autoregressive (NARX) neural network-type models with the same inputs as the corresponding SCAR-type models can lead to yet better performances. While individual Seasonal Component Artificial Neural Network (SCANN) models are generally worse than the corresponding SCAR-type structures, we provide empirical evidence that committee machines of SCANN networks can outperform the latter significantly.
-
Importance of the long-term Seasonal Component in day-ahead electricity price forecasting revisited: Neural network models
HSC Research Reports, 2017Co-Authors: Grzegorz Marcjasz, Bartosz Uniejewski, Rafał WeronAbstract:In day-ahead electricity price forecasting the daily and weekly Seasonalities are always taken into account, but the long-term Seasonal Component was believed to add unnecessary complexity and in most studies ignored. The recent introduction of the Seasonal Component AutoRegressive (SCAR) modeling framework has changed this viewpoint. However, the latter is based on linear models estimated using Ordinary Least Squares. Here we show that considering non-linear neural network-type models with the same inputs as the corresponding SCAR model can lead to a yet better performance. While individual Seasonal Component Artificial Neural Network (SCANN) models are generally worse than the corresponding SCAR-type structures, we provide empirical evidence that committee machines of SCANN networks can significantly outperform the latter.
-
On the importance of the long-term Seasonal Component in day-ahead electricity price forecasting. Part II – Probabilistic forecasting
HSC Research Reports, 2017Co-Authors: Bartosz Uniejewski, Grzegorz Marcjasz, Rafał WeronAbstract:A recent electricity price forecasting study has shown that the Seasonal Component AutoRegressive (SCAR) modeling framework, which consists of decomposing a series of spot prices into a trend-Seasonal and a stochastic Component, modeling them independently and then combining their forecasts, can yield more accurate point predictions than an approach in which the same autoregressive model is calibrated to the prices themselves. Here, we show that further accuracy gains can be achieved when the explanatory variables (load forecasts) are deSeasonalized as well. More importantly, considering a novel extension of the SCAR concept to probabilistic forecasting and applying two methods of combining predictive distributions we find that (i) SCAR-type models nearly always significantly outperform the autoregressive benchmark but are in turn outperformed by combined SCAR forecasts, (ii) predictive distributions computed using Quantile Regression Averaging (QRA) outperform those obtained from historical simulation and bootstrap methods, and (iii) averaging over predictive distributions generally yields better probabilistic forecasts of electricity spot prices than averaging over quantiles.
-
SCAR: MATLAB function to compute day-ahead predictions of the electricity spot price using the Seasonal Component AutoRegressive (SCAR) model
2016Co-Authors: Jakub Nowotarski, Rafał WeronAbstract:SCAR computes day-ahead predictions of the electricity spot price given historical prices and forecasted loads using the Seasonal Component AutoRegressive (SCAR) model introduced in J. Nowotarski, R. Weron (2016) "On the importance of the long-term Seasonal Component in day-ahead electricity price forecasting", Energy Economics 57, 228-235 (http://dx.doi.org/10.1016/j.eneco.2016.05.009), as well as the classical AR model.
Vilém Novák - One of the best experts on this subject based on the ideXlab platform.
-
Forecasting Seasonal time series based on fuzzy techniques
Fuzzy Sets and Systems, 2019Co-Authors: Linh Nguyen, Vilém NovákAbstract:Abstract This paper is devoted to a method for the forecasting of Seasonal time series. The core of our approach is based on the fuzzy transform and fuzzy natural logic (FNL) techniques. Under the assumption that a time series can be additively decomposed into a trend-cycle, a Seasonal Component and an irregular fluctuation, the forecasting is a combination of individual forecasting of each of these constituents. More precisely, the trend-cycle and the Seasonal Component are predicted with the help of fuzzy transform, pattern recognition and fuzzy natural logic techniques. To model the irregular Component, we apply the Box–Jenkins approach. In the paper, we compare the suggested method with two other well-known methods, namely STL (see [14] ) and ARIMA ones.
-
FUZZ-IEEE - Filtering out high frequencies in time series using F-transform with respect to raised cosine generalized uniform fuzzy partition
2015 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), 2015Co-Authors: Linh Nguyen, Vilém NovákAbstract:In this paper, we apply the F-transform technique with respect to raised cosine generalized uniform fuzzy partition to analysis of time series. We assume that the time series can be additively decomposed into trend-cycle, Seasonal Component and noise. We will prove that when constructing reasonably the fuzzy partition then the inverse F-transform of the time series provides a good approximation of its trend-cycle. Namely, the F-transform with respect to raised cosine generalized uniform fuzzy partition is able to eliminate completely the Seasonal Component and significantly reduce noise.
-
Filtering out high frequencies in time series using F-transform
Information Sciences, 2014Co-Authors: Vilém Novák, Irina Perfilieva, Michal Holčapek, Vladik KreinovichAbstract:Abstract In this paper, we focus on application of fuzzy transform (F-transform) to analysis of time series under the assumption that the latter can be additively decomposed into trend-cycle, Seasonal Component and noise. We prove that when setting properly width of the basic functions, the inverse F-transform of the time series closely approximates its trend-cycle. This means that the F-transform almost completely removes the Seasonal Component and noise. The obtained theoretical results are demonstrated on two artificial time series whose trend cycle is precisely known and on three real time series. At the same time, comparison with three classical methods, namely STL-method, SSA-method and low pass Butterworth filter is also provided.
Jakub Nowotarski - One of the best experts on this subject based on the ideXlab platform.
-
SCAR: MATLAB function to compute day-ahead predictions of the electricity spot price using the Seasonal Component AutoRegressive (SCAR) model
2016Co-Authors: Jakub Nowotarski, Rafał WeronAbstract:SCAR computes day-ahead predictions of the electricity spot price given historical prices and forecasted loads using the Seasonal Component AutoRegressive (SCAR) model introduced in J. Nowotarski, R. Weron (2016) "On the importance of the long-term Seasonal Component in day-ahead electricity price forecasting", Energy Economics 57, 228-235 (http://dx.doi.org/10.1016/j.eneco.2016.05.009), as well as the classical AR model.
-
SCAR_EXAMPLE: MATLAB codes and data for "On the importance of the long-term Seasonal Component in day-ahead electricity price forecasting"
2016Co-Authors: Jakub Nowotarski, Rafał WeronAbstract:This Matlab script illustrates the concepts and methods presented in J. Nowotarski, R. Weron (2016) "On the importance of the long-term Seasonal Component in day-ahead electricity price forecasting", Energy Economics 57, 228-235 (http://dx.doi.org/10.1016/j.eneco.2016.05.009) and together with the accompanying codes and data files can be used to replicate the Figures and Tables in the text. To obtain all entries in Table 1, change arxFlag, seasFlag and smoothing parameters. The function scar.m can be used to calibrate both the newly introduced Seasonal Component AutoRegressive class of models (SCAR, SCARX), as well as the classical AR and ARX models. The zip file includes one script (main.m), one m-file (scar.m) and three data files.
-
On the importance of the long-term Seasonal Component in day-ahead electricity price forecasting
Energy Economics, 2016Co-Authors: Jakub Nowotarski, Rafał WeronAbstract:In day-ahead electricity price forecasting (EPF) the daily and weekly Seasonalities are always taken into account, but the long-term Seasonal Component (LTSC) is believed to add unnecessary complexity to the already parameter-rich models and is generally ignored. Conducting an extensive empirical study involving state-of-the-art time series models we show that (i) decomposing a series of electricity prices into a LTSC and a stochastic Component, (ii) modeling them independently and (iii) combining their forecasts can bring – contrary to a common belief – an accuracy gain compared to an approach in which a given time series model is calibrated to the prices themselves.
-
LTSC_EXAMPLE: MATLAB example script and data for "Robust estimation and forecasting of the long-term Seasonal Component of electricity spot prices"
2013Co-Authors: Jakub Nowotarski, Rafał WeronAbstract:This Matlab script illustrates the concepts and methods presented in J.Nowotarski, J.Tomczyk, R.Weron (2013) "Robust estimation and forecasting of the long-term Seasonal Component of electricity spot prices", Energy Economics 39, 13-27 (http://dx.doi.org/10.1016/j.eneco.2013.04.004). The zip file includes one script and one data file. The script requires 3 Matlab functions for proper operation: ltscsimple.m, ltscsin.m and ltscwave.m available from http://ideas.repec.org/s/wuu/hscode.html.
-
LTSCWAVE: MATLAB function to estimate and forecast the long-term Seasonal Component (LTSC) of an electricity spot price series using wavelet-based methods
2013Co-Authors: Jakub Nowotarski, Jakub Tomczyk, Rafał WeronAbstract:LTSCWAVE returns the long-term Seasonal Component (LTSC) of an electricity spot price series and its forecast up to a year ahead computed using one of the wavelet-based methods studied in J. Nowotarski, J. Tomczyk, R. Weron (2013), "Robust estimation and forecasting of the long-term Seasonal Component of electricity spot prices", Energy Economics, forthcoming (doi:10.1016/j.eneco.2013.04.004).
Linh Nguyen - One of the best experts on this subject based on the ideXlab platform.
-
Forecasting Seasonal time series based on fuzzy techniques
Fuzzy Sets and Systems, 2019Co-Authors: Linh Nguyen, Vilém NovákAbstract:Abstract This paper is devoted to a method for the forecasting of Seasonal time series. The core of our approach is based on the fuzzy transform and fuzzy natural logic (FNL) techniques. Under the assumption that a time series can be additively decomposed into a trend-cycle, a Seasonal Component and an irregular fluctuation, the forecasting is a combination of individual forecasting of each of these constituents. More precisely, the trend-cycle and the Seasonal Component are predicted with the help of fuzzy transform, pattern recognition and fuzzy natural logic techniques. To model the irregular Component, we apply the Box–Jenkins approach. In the paper, we compare the suggested method with two other well-known methods, namely STL (see [14] ) and ARIMA ones.
-
TREND-CYCLE ESTIMATION USING FUZZY TRANSFORM OF HIGHER DEGREE
Iranian Journal of Fuzzy Systems, 2018Co-Authors: Michal Holčapek, Linh NguyenAbstract:In this paper, we provide theoretical justification for the application of higher degree fuzzy transform in time series analysis. Under the assumption that a time series can be additively decomposed into a trend-cycle, a Seasonal Component and a random noise, we demonstrate that the higher degree fuzzy transform technique can be used for the estimation of the trend-cycle, which is one of the basic tasks in time series analysis. We prove that high frequencies appearing in the Seasonal Component can be arbitrarily suppressed and that random noise, as a stationary process, can be successfully decreased using the fuzzy transform of higher degree with a reasonable adjustment of parameters of a generalized uniform fuzzy partition.
-
IPMU (2) - Suppression of High Frequencies in Time Series Using Fuzzy Transform of Higher Degree
Information Processing and Management of Uncertainty in Knowledge-Based Systems, 2016Co-Authors: Michal Holčapek, Linh NguyenAbstract:In this paper, we provide a theoretical justification for the application of higher degree fuzzy transform in time series analysis. We demonstrate that the higher degree fuzzy transform technique can be used for the suppression of high frequencies in time series, which belongs among the essential assumptions for a successful extraction of the trend (trend-cycle) of time series. More precisely, if a time series can be additively decomposed into a trend-cycle, a Seasonal Component and a noise, we show that high frequencies appearing in the Seasonal Component can be arbitrarily suppressed using the fuzzy transform of higher degree with a reasonable adjustment of parameters of a generalized uniform fuzzy partition.
-
FUZZ-IEEE - Filtering out high frequencies in time series using F-transform with respect to raised cosine generalized uniform fuzzy partition
2015 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), 2015Co-Authors: Linh Nguyen, Vilém NovákAbstract:In this paper, we apply the F-transform technique with respect to raised cosine generalized uniform fuzzy partition to analysis of time series. We assume that the time series can be additively decomposed into trend-cycle, Seasonal Component and noise. We will prove that when constructing reasonably the fuzzy partition then the inverse F-transform of the time series provides a good approximation of its trend-cycle. Namely, the F-transform with respect to raised cosine generalized uniform fuzzy partition is able to eliminate completely the Seasonal Component and significantly reduce noise.
Rafal Weron - One of the best experts on this subject based on the ideXlab platform.
-
on the importance of the long term Seasonal Component in day ahead electricity price forecasting part ii probabilistic forecasting
Energy Economics, 2019Co-Authors: Bartosz Uniejewski, Grzegorz Marcjasz, Rafal WeronAbstract:A recent electricity price forecasting study has shown that the Seasonal Component AutoRegressive (SCAR) modeling framework, which consists of decomposing a series of spot prices into a trend-Seasonal and a stochastic Component, modeling them independently and then combining their forecasts, can yield more accurate point predictions than an approach in which the same autoregressive model is calibrated to the prices themselves. Here, we show that further accuracy gains can be achieved when the explanatory variables (load forecasts) are deSeasonalized as well. More importantly, considering a novel extension of the SCAR concept to probabilistic forecasting and applying two methods of combining predictive distributions, we find that (i) SCAR-type models nearly always significantly outperform the autoregressive benchmark but are in turn outperformed by combined SCAR forecasts, (ii) predictive distributions computed using Quantile Regression Averaging (QRA) outperform those obtained from historical simulation and bootstrap methods, and (iii) averaging over predictive distributions generally yields better probabilistic forecasts of electricity spot prices than averaging over quantiles. Given that probabilistic forecasting is a concept closely related to risk management, our study has important implications for risk officers and portfolio managers in the power sector.
-
on the importance of the long term Seasonal Component in day ahead electricity price forecasting part ii probabilistic forecasting
HSC Research Reports, 2017Co-Authors: Bartosz Uniejewski, Grzegorz Marcjasz, Rafal WeronAbstract:A recent electricity price forecasting study has shown that the Seasonal Component AutoRegressive (SCAR) modeling framework, which consists of decomposing a series of spot prices into a trend-Seasonal and a stochastic Component, modeling them independently and then combining their forecasts, can yield more accurate point predictions than an approach in which the same autoregressive model is calibrated to the prices themselves. Here, we show that further accuracy gains can be achieved when the explanatory variables (load forecasts) are deSeasonalized as well. More importantly, considering a novel extension of the SCAR concept to probabilistic forecasting and applying two methods of combining predictive distributions we find that (i) SCAR-type models nearly always significantly outperform the autoregressive benchmark but are in turn outperformed by combined SCAR forecasts, (ii) predictive distributions computed using Quantile Regression Averaging (QRA) outperform those obtained from historical simulation and bootstrap methods, and (iii) averaging over predictive distributions generally yields better probabilistic forecasts of electricity spot prices than averaging over quantiles.
-
deSeasonalize matlab function to remove short and long term Seasonal Components new implementation
HSC Software, 2014Co-Authors: Rafal WeronAbstract:[DDATA,LTSC,STSC] = DESeasonalIZE(DATA,MET1,MET2,HOLI) returns deSeasonalized data vector DDATA, long-term Seasonal Component LTSC and short-term Seasonal Component STSC obtained from the original data series DATA. Parameters MET1 and MET2 define the short- and long-term Seasonal decomposition techniques, respectively. HOLI is a vector of day types, e.g. [1,2,3,4,5,6,7,8] where 8 stands for holidays. This is version 2015-02-19.
-
Identifying Spikes and Seasonal Components in Electricity Spot Price Data: A Guide to Robust Modeling
SSRN Electronic Journal, 2012Co-Authors: Joanna Janczura, Rafal Weron, Stefan Trueck, Rodney WolffAbstract:We investigate the effects of outlier treatment on the estimation of the Seasonal Component and stochastic models in electricity markets. Typically, electricity spot prices exhibit features like Seasonality, mean-reverting behavior, extreme volatility and the occurrence of jumps and spikes. Hence, an important issue in the estimation of stochastic models for electricity spot prices is the estimation of a Component to deal with trends and Seasonality in the data. Unfortunately, in regression analysis, classical estimation routines like OLS are very sensitive to extreme observations and outliers. Improved robustness of the model can be achieved by (a) cleaning the data with some reasonable procedure for outlier rejection, and then (b) using classical estimation and testing procedures on the remainder of the data. We examine the effects on model estimation for different treatment of extreme observations in particular on determining the number of outliers and descriptive statistics of the remaining series after replacement of the outliers. Our findings point out the substantial impact the treatment of extreme observations may have on these issues.
-
remst matlab function to remove trend and Seasonal Component using the moving average method
Statistical Software Components, 2010Co-Authors: Rafal WeronAbstract:Y = REMST returns a time series with removed polynomial trend and Seasonal Components of a given period. As additional output parameters it also returns the identified Seasonal Component and the fitted polynomial coefficients. REMST uses the moving average technique (see eg. Weron (2006) "Modeling and Forecasting Electricity Loads and Prices", Wiley, Section 2.4.3).
