Assessment of stacked unidirectional and bidirectional long short-term memory networks for electricity load forecasting

Atef, Sara; Eltawil, Amr B.

doi:10.1016/j.epsr.2020.106489

Cited by 73 publications

(39 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The best model occurs in S6, which means that the bi-directional architecture performs better than the single directional architecture for two-day-ahead forecasting after hyperparameter optimisation. The outperformance of the bi-directional architecture has been claimed by many studies [52,53,60,66] in other domains and is still supported by this research when applying to runoff forecasting.…”

Section: Overall Evaluationsupporting

confidence: 73%

“…The advantage of the PCA when applying to the preprocessing stage in machine learning is also supported by studies in other realms [56,58,70,71]. Similar to the results of the research about NLP [36,72] and other topics [53,73,74], the bi-directional RNN architecture also tends to provide more accurate results and is recommended to be utilized in the GRU runoff forecasting model. For further studies, more data filtering strategies can be tested.…”

Section: Recommendations Based On the Evaluation Resultsmentioning

confidence: 56%

“…The combination of several layers of the basic LSTM or GRU cells may enhance the model's ability when learning the nonlinear relationships and provide a better prediction accuracy [53,60]. The typical structure of a stacked GRU network with two layers is shown in Figure 3.…”

Section: Stacked and Bi-directional Gru (Bi-gru)mentioning

confidence: 99%

“…Another study [51] shows that a better prediction result (the test NSE of 0.942 compared with 0.666) has been obtained when using a selected subset of all rain gauges as the input. Furthermore, the LSTM or GRU networks' architectures can also influence the forecasting results [52,53].…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Impact of Input Filtering and Architecture Selection Strategies on GRU Runoff Forecasting: A Case Study in the Wei River Basin, Shaanxi, China

Wang

Liu

Yue

et al. 2020

Water

View full text Add to dashboard Cite

A gated recurrent unit (GRU) network, which is a kind of artificial neural network (ANN), has been increasingly applied to runoff forecasting. However, knowledge about the impact of different input data filtering strategies and the implications of different architectures on the GRU runoff forecasting model’s performance is still insufficient. This study has selected the daily rainfall and runoff data from 2007 to 2014 in the Wei River basin in Shaanxi, China, and assessed six different scenarios to explore the patterns of that impact. In the scenarios, four manually-selected rainfall or runoff data combinations and principal component analysis (PCA) denoised input have been considered along with single directional and bi-directional GRU network architectures. The performance has been evaluated from the aspect of robustness to 48 various hypermeter combinations, also, optimized accuracy in one-day-ahead (T + 1) and two-day-ahead (T + 2) forecasting for the overall forecasting process and the flood peak forecasts. The results suggest that the rainfall data can enhance the robustness of the model, especially in T + 2 forecasting. Additionally, it slightly introduces noise and affects the optimized prediction accuracy in T + 1 forecasting, but significantly improves the accuracy in T + 2 forecasting. Though with relevance (R = 0.409~0.763, Grey correlation grade >0.99), the runoff data at the adjacent tributary has an adverse effect on the robustness, but can enhance the accuracy of the flood peak forecasts with a short lead time. The models with PCA denoised input has an equivalent, even better performance on the robustness and accuracy compared with the models with the well manually filtered data; though slightly reduces the time-step robustness, the bi-directional architecture can enhance the prediction accuracy. All the scenarios provide acceptable forecasting results (NSE of 0.927~0.951 for T + 1 forecasting and 0.745~0.836 for T + 2 forecasting) when the hyperparameters have already been optimized. Based on the results, recommendations have been provided for the construction of the GRU runoff forecasting model.

show abstract

Section: Overall Evaluationsupporting

confidence: 73%

Section: Recommendations Based On the Evaluation Resultsmentioning

confidence: 56%

Section: Stacked and Bi-directional Gru (Bi-gru)mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Impact of Input Filtering and Architecture Selection Strategies on GRU Runoff Forecasting: A Case Study in the Wei River Basin, Shaanxi, China

Wang

Liu

Yue

et al. 2020

Water

View full text Add to dashboard Cite

show abstract

“…Results show bidirectional model helps to further improve prediction accuracy [23]. Atef et al conduct a systematic experimental methodology to investigate the impact of using deep-stacked unidirectional and bidirectional networks on predicting electricity load consumption and draw a conclusion that bidirectional models have significant improvement in the prediction accuracy while they consume almost twice the time of the unidirectional models [24]. However, due to the doubling of the learn-able parameters, the training time of Bi-RNN-based models is also doubled, that is, the efficiency is greatly reduced.…”

Section: Literature Reviewmentioning

confidence: 99%

Short-Term Load Forecasting Using Recurrent Neural Networks With Input Attention Mechanism and Hidden Connection Mechanism

Zhang

2020

IEEE Access

View full text Add to dashboard Cite

Short-term load forecasting is a critical task in the smart grid, which can be used to optimize power deployment and reduce power losses. Recurrent neural networks (RNNs) are the most popular deep learning models for short-term load forecasting. However, despite of achieving better forecasting accuracy than the traditional models, the performance of the existing RNN-based load forecasting approaches is still unsatisfactory for practical usage. Therefore, in this work, we have proposed input attention mechanism (IAM) and hidden connection mechanism (HCM) to greatly enhance the accuracy and efficiency of RNNbased load forecasting models. Specifically, we use IAM to assign the importance weights on input layers, which have better performances in both efficiency and accuracy than traditional attention mechanisms. To further enhance the models' efficiency, HCM is then applied to utilize residual connections to enhance the model's converging speed. We have applied both IAM and HCM on four state-of-the-art RNN implementations, and then conducted extensive experimental studies on two public datasets. Experimental results show that the proposed RNNs with IAM and HCM models achieve much better performances than the state-of-the-art baselines in both accuracy and efficiency. Ablation studies show that both IAM and HCM are essential to achieve such superior performances.

show abstract

Comparison of the advanced machine learning methods for better prediction accuracy of solar radiation using only temperature data: A case study

Mirbolouki

Heddam

Parmar

et al. 2021

Intl J of Energy Research

View full text Add to dashboard Cite

Estimation of solar radiation (SR) carries importance for planning available renewable energy, and it is also beneficial for solving agricultural, meteorological, and engineering problems. This study compares the ability of hybrid adaptive neuro fuzzy (ANFIS) models and long short-term memory to search a suitable approach for SR prediction with minimum number of input parameters (temperature) in Mediterranean region of Turkey, which could be useful for the regions in which other effective parameters (eg, relative humidity, wind speed) are not available. The models considered were assessed by considering four data splitting scenarios, 50% train-50% test, 60% train-40% test, 70% train-30% test, and 80% train-20% test. Among the hybrid methods, the ANFIS with grey wolf optimization and genetic algorithm showed a superior accuracy. The study shows that applying different data splitting scenarios is necessary for better assessment of the data-driven methods since the accuracies of the implemented methods increase by about 30% to 60% when the splitting data scenario varies from 50-50% to 80-20%. Sensitivity analysis shows that the performance of the model increases by about 40% using extraterrestrial radiation for the best model. The ANFIS with grey wolf optimization and genetic algorithm is recommended to predict monthly solar radiation with limited input data.

show abstract

Assessment of stacked unidirectional and bidirectional long short-term memory networks for electricity load forecasting

Cited by 73 publications

References 36 publications

Impact of Input Filtering and Architecture Selection Strategies on GRU Runoff Forecasting: A Case Study in the Wei River Basin, Shaanxi, China

Impact of Input Filtering and Architecture Selection Strategies on GRU Runoff Forecasting: A Case Study in the Wei River Basin, Shaanxi, China

Short-Term Load Forecasting Using Recurrent Neural Networks With Input Attention Mechanism and Hidden Connection Mechanism

Comparison of the advanced machine learning methods for better prediction accuracy of solar radiation using only temperature data: A case study

Contact Info

Product

Resources

About