Bayesian deep learning based method for probabilistic forecast of day-ahead electricity prices

Brusaferri, Alessandro; Matteucci, Matteo; Portolani, Pietro; Vitali, Andrea

doi:10.1016/j.apenergy.2019.05.068

Cited by 107 publications

(40 citation statements)

References 80 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Besides, sensible shifts occur between working days and holidays, mostly due to different consumption patterns. A more detailed analysis of the Italian day-ahead market dataset, including descriptive statistics, can be found in [7].…”

Section: Resultsmentioning

confidence: 99%

“…To this end, batches of ordered sub-sequences (including both past price values and related conditioning variables) are built by sliding a window with configurable width throughout the overall sequences. Considering the analysis performed in [7], Fig. 3.…”

Section: Resultsmentioning

confidence: 99%

“…The proposed EPF approach has been compared to the Bayesian Neural Network (BNN) based method proposed in [7] and to a deterministic RNN based model. We considered the former since it represents, to the best of author knowledge, the unique probabilistic EPF method exploiting a Neural Network in the literature.…”

Section: Resultsmentioning

confidence: 99%

“…However, a probabilistic model capable of providing insight into the uncertainty in the prediction can be an invaluable tool both for generation companies and large-scale consumers who deal with the various source of stochasticity in their processes and need to account for risk in the decision-making chain. To the best of our knowledge, the only previous work exploring modern deep neural networks in a probabilistic framework is [7], but still in a simplified Gaussian distribution assumption. Indeed, as stated in the recent review [1], probabilistic EPF is a "fascinating but still underdeveloped" field.…”

Section: Context and Motivationmentioning

confidence: 99%

See 3 more Smart Citations

Probabilistic day-ahead energy price forecast by a Mixture Density Recurrent Neural Network

Brusaferri

Matteucci

Ramin³

et al. 2020

2020 7th International Conference on Control, Decision and Information Technologies (CoDIT)

Self Cite

View full text Add to dashboard Cite

Probabilistic electricity price forecast (EPF) systems represent a fundamental tool to achieve robust production scheduling and day-ahead bidding strategies. However, most EPF methods, including recently proposed deep learning based techniques, are still targeting point predictions, following the common Gaussian assumption. In this work, we propose a novel probabilistic EPF approach based on the integration of a Gaussian Mixture layer, parametrized by a Recurrent Neural Network with Gated Recurrent Units, including an L1-norm based feature selection mechanisms. The network is conceived to approximate general conditional price distributions through learning. Moreover, we developed a multi-hours prediction approach exploiting correlations and patters both in hourly and cross-hour contexts. Experiments have been performed on the Italian market dataset, showing the capability of the proposed method to achieve accurate out-of-sample predictions while providing explicit uncertainty indications supporting enhanced decision making.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

Section: Context and Motivationmentioning

confidence: 99%

See 2 more Smart Citations

Probabilistic day-ahead energy price forecast by a Mixture Density Recurrent Neural Network

Brusaferri

Matteucci

Ramin³

et al. 2020

2020 7th International Conference on Control, Decision and Information Technologies (CoDIT)

Self Cite

View full text Add to dashboard Cite

show abstract

“…To date, deep learning and meta ensemble learning have been widely applied in the fields of classification, pattern recognition and prediction. A new interval energy price prediction model was proposed based on Bayesian deep learning, and a new training method was designed for the model [24]. The study in [25] proposed a data-driven prediction program using deep learning to predict the remaining battery life.…”

Section: Introductionmentioning

confidence: 99%

Air-Conditioning Load Forecasting for Prosumer Based on Meta Ensemble Learning

Chen¹,

2020

IEEE Access

View full text Add to dashboard Cite

Accurate and reliable prediction of airconditioning load plays a significant role in prosumer energy management system (EMS), because airconditioning load accounts for a large proportion of the building's total energy consumption. This paper proposes a new meta ensemble learning method to realize short-term prediction of airconditioning load for prosumers. This method is a hybrid of meta ensemble learning and stacked auto-encoder (SAE). First, we design multiple different forecasting structures based on SAE to achieve point prediction of airconditioning loads. SAE is used to learn the deep features in airconditioning load data. Second, a new meta ensemble learning prediction model is proposed. Meta ensemble learning is used to learn the nonlinear features and invariant structures in data, and determine the coefficients of each SAE-based point forecaster. Finally, the prediction results of each point forecaster are aggregated and integrated to estimate the final airconditioning load prediction result. Airconditioning load data from a commercial building in Singapore are used to validate the feasibility and effectiveness of the proposed method, demonstrating that the proposed meta ensemble learning method is attractive in prosumer energy management. INDEX TERMS Airconditioning load, load prediction, meta learning, ensemble learning, prosumer energy management system.

show abstract