Ensemble Empirical Mode Decomposition Based Deep Learning Model for Short-Term Wind Power Forecasting

2022 ◽  
Author(s):  
J.M. González-Sopeña

Abstract. In the last few years, wind power forecasting has established itself as an essential tool in the energy industry due to the increase of wind power penetration in the electric grid. This paper presents a wind power forecasting method based on ensemble empirical mode decomposition (EEMD) and deep learning. EEMD is employed to decompose wind power time series data into several intrinsic mode functions and a residual component. Afterwards, every intrinsic mode function is trained by means of a CNN-LSTM architecture. Finally, wind power forecast is obtained by adding the prediction of every component. Compared to the benchmark model, the proposed approach provides more accurate predictions for several time horizons. Furthermore, prediction intervals are modelled using quantile regression.

2013 ◽  
Vol 2 (2) ◽  
pp. 192-199 ◽  
Author(s):  
Zeng-Wei Zheng ◽  
Yuan-Yi Chen ◽  
Xiao-Wei Zhou ◽  
Mei-Mei Huo ◽  
Bo Zhao ◽  
...  

Author(s):  
Weifei Hu ◽  
Yihan He ◽  
Zhenyu Liu ◽  
Jianrong Tan ◽  
Ming Yang ◽  
...  

Abstract Precise time series prediction serves as an important role in constructing a Digital Twin (DT). The various internal and external interferences result in highly non-linear and stochastic time series data sampled from real situations. Although artificial Neural Networks (ANNs) are often used to forecast time series for their strong self-learning and nonlinear fitting capabilities, it is a challenging and time-consuming task to obtain the optimal ANN architecture. This paper proposes a hybrid time series prediction model based on ensemble empirical mode decomposition (EEMD), long short-term memory (LSTM) neural networks, and Bayesian optimization (BO). To improve the predictability of stochastic and nonstationary time series, the EEMD method is implemented to decompose the original time series into several components, each of which is composed of single-frequency and stationary signal, and a residual signal. The decomposed signals are used to train the BO-LSTM neural networks, in which the hyper-parameters of the LSTM neural networks are fine-tuned by the BO algorithm. The following time series data are predicted by summating all the predictions of the decomposed signals based on the trained neural networks. To evaluate the performance of the proposed hybrid method (EEMD-BO-LSTM), this paper conducts a case study of wind speed time series prediction and has a comprehensive comparison between the proposed method and other approaches including the persistence model, ARIMA, LSTM neural networks, B0-LSTM neural networks, and EEMD-LSTM neural networks. Results show an improved prediction accuracy using the EEMD-BO-LSTM method by multiple accuracy metrics.


2020 ◽  
Vol 19 (Number 4) ◽  
pp. 533-558
Author(s):  
Mohammad Raquibul Hossain ◽  
Mohd Tahir Ismail

Forecasting is a challenging task as time series data exhibit many features that cannot be captured by a single model. Therefore, many researchers have proposed various hybrid models in order to accommodate these features to improve forecasting results. This work proposed a hybrid method between Empirical Mode Decomposition (EMD) and Theta methods by considering better forecasting potentiality. Both EMD and Theta are efficient methods in their own ground of tasks for decomposition and forecasting, respectively. Combining them to obtain a better synergic outcome deserves consideration. EMD decomposed the training data from each of the five Financial Times Stock Exchange 100 Index (FTSE 100 Index) companies’ stock price time series data into Intrinsic Mode Functions (IMF) and residue. Then, the Theta method forecasted each decomposed subseries. Considering different forecast horizons, the effectiveness of this hybridisation was evaluated through values of conventional error measures found for test data and forecast data, which were obtained by adding forecast results for all component counterparts extracted from the EMD process. This study found that the proposed method produced better forecast accuracy than the other three classic methods and the hybrid EMD-ARIMA models.


2006 ◽  
Vol 2006 ◽  
pp. 1-17 ◽  
Author(s):  
Md. Khademul Islam Molla ◽  
M. Sayedur Rahman ◽  
Akimasa Sumi ◽  
Pabitra Banik

We have used empirical mode decomposition (EMD) method, which is especially well fitted for analyzing time-series data representing nonstationary and nonlinear processes. This method could decompose any time-varying data into a finite set of functions called “intrinsic mode functions” (IMFs). The EMD analysis successively extracts the IMFs with the highest local temporal frequencies in a recursive way. The extracted IMFs represent a set of successive low-pass spatial filters based entirely on the properties exhibited by the data. The IMFs are mutually orthogonal and more effective in isolating physical processes of various time scales. The results showed that most of the IMFs have normal distribution. Therefore, the energy density distribution of IMF samples satisfiesχ2-distribution which is statistically significant. This study suggested that the recent global warming along with decadal climate variability contributes not only to the more extreme warm events, but also to more frequent, long lasting drought and flood.


2021 ◽  
Vol 10 (2) ◽  
pp. 211-220
Author(s):  
Rosinar Siregar ◽  
Rukun Santoso ◽  
Puspita Kartikasari

 Stock price fluctuations make investors tend to hesitate to invest in stock markets because of an uncertain situation in the future. One method that can solve these problems is to use forecasting about the stock prices in the future. Generally, the huge size of data non linear and non stationary, and it is difficult to be interpreted in concrete. This problem can be solved by performing the decomposition process. One of decomposition method in time series data is Ensemble Empirical Mode Decomposition (EEMD). EEMD is process decomposition data into several Intrinsic Mode Function (IMF) and the IMF residue. In this research, this concept applied to data Stock Price Index in Property, Real Estate, and Construction from July 1, 2019 to July 30, 2020 as many as 272 data. Based on the results of data processing, as many as 6 IMF and IMF remaining were used as IMF forecasting and the IMF remaining in the future. The forecast was performed by choosing the best model of each IMF component and IMF remaining, used ARIMA and polynomial trend. Keywords: Time Series Data, Stock Price Index, EEMD, ARIMA, Polynomial Trend.


Sign in / Sign up

Export Citation Format

Share Document