DA-LSTM: A dynamic drift-adaptive learning framework for interval load forecasting with LSTM networks

Bayram, Firas; Aupke, Phil; Ahmed, Bestoun S.; Kassler, Andreas; Theocharis, Andreas; Forsman, Jukka T.

doi:10.1016/j.engappai.2023.106480

Cited by 23 publications

(5 citation statements)

References 76 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The results show that this model has the best performance with a MAPE of 8.26% and an RMSE of 21.41% when tested using the local CPU and GPU on the device and the CPU and GPU on the cloud-web service (AWS). This study demonstrates the potential and reliability of the LSTM model for use in predicting electrical loads [11].…”

Section: Introductionmentioning

confidence: 65%

Comparison of KNN and LSTM on the Prediction of the Operational Conditions of Natural Gas Pipeline Transmission Networks

Yusuf,

Hasmawati,

Ihsan

2024

jipi. jurnal. ilmiah. penelitian. dan. pembelajaran. informatik

View full text Add to dashboard Cite

During the gas distribution process, a sequence of compressors creates a pressure difference, causing gas to move from regions of high pressure to areas with comparatively lower pressure. The Natural Gas transmission process experiences variations in pressure and temperature, primarily caused by frictional losses, differences in altitude, gas velocity, and the Joule-Thompson effect. Additionally, effective heat transfer to or from the environment contributes to temperature changes throughout the pipeline. The presence of liquid and density changes (hydrate) within the channel also has an impact on the pressure, influencing both pressure and temperature conditions.. This study implements the KNN and LSTM models to predict pressure conditions in natural gas transmission pipelines to analyze the performance comparison of the best model performance using several appropriate parameters to support maximum method performance results. The results show that the LSTM model is better at predicting pressure conditions in natural gas pipeline transmission networks, with an R2 score of 99.45, compared to the KNN model, with an R2 score of 92.82. This study also obtained prediction results from the KNN and LSTM models; the KNN model tends to produce the same pressure value for eight months, while the LSTM model produces pressure values that tend to vary.

show abstract

Section: Introductionmentioning

confidence: 65%

Comparison of KNN and LSTM on the Prediction of the Operational Conditions of Natural Gas Pipeline Transmission Networks

Yusuf,

Hasmawati,

Ihsan

2024

jipi. jurnal. ilmiah. penelitian. dan. pembelajaran. informatik

View full text Add to dashboard Cite

show abstract

“…Therefore, a model adaptation technique is required to detect model drift based on a prespecified decision threshold and to maintain performance. The adaptation approach usually involves a model update step that reidentifies the coefficients of the previous model based on new data information. , In this study, we set the cumulative absolute percentage error for the five latest past predictions (CAPE 5 ) as an adaptation trigger to guide decision-making regarding the model update, and its mathematical expression is as follows CAP normalE 5 = prefix∑ i = t − 4 t | false| Y i − Ŷ i false| | F | false| Y i false| | F where Y i and Ŷ i are the mean values of the actual process data and predicted data from the regression model, respectively. The Frobenius norm ( | false| Y i false| | F = ∑ i = 1 m ∑ j = 1 n false| a ij false| 2 ) is a norm of matrix A, defined as the square root of the sum of the absolute squares of its elements a ij .…”

Section: Methodsmentioning

confidence: 99%

“…The adaptation approach usually involves a model update step that reidentifies the coefficients of the previous model based on new data information. 41,42 In this study, we set the cumulative absolute percentage error for the five latest past predictions (CAPE 5 ) as an adaptation trigger to guide decision-making regarding the model update, and its mathematical expression is as follows…”

Section: Developed Adaptive Regression Modeling Methodologymentioning

confidence: 99%

Machine Learning-Based Adaptive Regression to Identify Nonlinear Dynamics of Biochemical Systems: A Case Study on Bio 2,3-Butanediol Distillation Process

Choi,

Bhadriraju,

Kwon

et al. 2024

ACS Sustainable Chem. Eng.

View full text Add to dashboard Cite

Developing an accurate process model is essential to efficiently operate a process and maximize its economics. While offline data-driven models utilizing historical data generally exhibit satisfactory performance, their effectiveness diminishes in accurately predicting real processes characterized by constant changes and uncertainties over time. Hence, there is a need for an adaptive model that is capable of effectively handling dynamic behavior. In this study, we propose an adaptive data-driven regression model that leverages subset selection techniques and decision thresholds. In addition, a comprehensive analysis was performed to determine the best adaptive regression model, considering case studies with different model parameters and training window sizes, taking into account statistical indicators of model accuracy as well as nonstatistical indicators such as the number of updates, update period, and computation time. The developed adaptive regression model has been successfully demonstrated on a bio 2,3-Butanediol distillation column at GS Caltex, Republic of Korea, suggesting its potential applicability to similar process systems and providing opportunities for future research in process optimization and control.

show abstract

“…Additionally, the architecture and structure of the dynamic DL model are updated using the moving horizon approach, allowing it to capture the most recent highlighting patterns in the building's electrical load. [14] can enhance the performance of load forecasting models without the need for a drift threshold. They incorporate a number of active and passive adaption mechanisms into the framework.…”

Section: Related Work (İlgi̇li̇ çAlişmalar)mentioning

confidence: 99%

A Deep Learning Approach to Real-time Electricity Load Forecasting

AL-HAMİD,

SAVAŞ

2023

Journal of Information Systems and Management Research

View full text Add to dashboard Cite

In light of the increasing importance of accurate and real-time electrical demand forecasting, this research presents a deep learning model with the goal of dramatically improving predictive accuracy. Conventional methods of forecasting, such as linear regression, have trouble capturing the complex patterns included in data about electricity usage. Standard machine learning methods are shown to be wanting when compared to the suggested deep Long Short-Term Memory (LSTM) model. Mean Absolute Error (MAE) of 5.454 and Mean Squared Error (MSE) of 18.243 demonstrate the deep LSTM model's proficiency in tackling this problem. The linear regression, on the other hand, achieved a MAE of 47.352 and an MSE of 65.606, which is lower than the proposed model. Because of its greater predictive precision and reliability, the deep LSTM model is a viable option for accurate, real-time prediction of electricity demand.

show abstract

DA-LSTM: A dynamic drift-adaptive learning framework for interval load forecasting with LSTM networks

Cited by 23 publications

References 76 publications

Comparison of KNN and LSTM on the Prediction of the Operational Conditions of Natural Gas Pipeline Transmission Networks

Comparison of KNN and LSTM on the Prediction of the Operational Conditions of Natural Gas Pipeline Transmission Networks

Machine Learning-Based Adaptive Regression to Identify Nonlinear Dynamics of Biochemical Systems: A Case Study on Bio 2,3-Butanediol Distillation Process

A Deep Learning Approach to Real-time Electricity Load Forecasting

Contact Info

Product

Resources

About