Continuous Learning of Deep Neural Networks to Improve Forecasts for Regional Energy Markets

He, Yujiang; Henze, Janosch; Sick, Bernhard

doi:10.1016/j.ifacol.2020.12.1017

Cited by 11 publications

(3 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…ANNs as part of AI-based methods have gained prominence (Falatouri et al 2022), often outperforming conventional approaches, as they can learn data-driven models connect-ing input and expected forecast data (LeCun et al 2015). ANN-based models can also be updated with additional training on newly connected data without repeating the entire learning process (He et al 2020). ANN-based models require a substantial amount of historical data, and not all the process data is available to researchers (Chen, Lin 2014).…”

Section: Literature Reviewmentioning

confidence: 99%

Freight Rate and Demand Forecasting in Road Freight Transportation Using Econometric and Artificial Intelligence Methods

Liachovičius,

Šabanovič,

Skrickij

2023

Transport

View full text Add to dashboard Cite

The digitisation of the transportation sector and data availability have opened up new opportunities to implement data-driven methods for improving company performance. This article analyses demand and freight rate forecasting techniques in the context of the road freight transportation company. The European market was analysed in this research, and direction from the Netherlands to Italy was selected for the case study. Performed investigation showed that econometric models such as Auto-Regressive Integrated Moving Average (ARIMA) used for demand prognosis provide good results. Freight rate forecasting is different; econometric models, including multivariate models ARIMA with exogenous variables (ARIMAX) and Seasonal ARIMAX (SARIMAX), do not perform satisfactorily under specified time intervals, therefore MultiLayer Perceptron (MLP) was used as a solution. It can be seen that Artificial Intelligence (AI) based methods provide better results. Despite its success, the AI-based approach alone is not recommended for practical implementation since forecasted input parameters are necessary. Lastly, the study uncovers a valuable insight. A strong correlation (0.86) between spot and contract rates was found, and the article shows how current spot rates can be used for contract rate forecasting.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Freight Rate and Demand Forecasting in Road Freight Transportation Using Econometric and Artificial Intelligence Methods

Liachovičius,

Šabanovič,

Skrickij

2023

Transport

View full text Add to dashboard Cite

show abstract

“…All of the above experiments address only classification problems. In [22], He et al propose two CL application scenarios for establishing regional smart grids: the task-domain incremental scenario and the datadomain incremental scenario. The scenarios are applicable for forecasting power, including renewable energy generation and power consumption in the middle-/lowvoltage grid.…”

Section: Related Workmentioning

confidence: 99%

“…Fitting error indicates how well the instance fits all seen samples after the update phase. Because a CLeaR instance is updated by mini-batch data every time, some batches may lead it to a local minimum during Forgetting ratio is a metric proposed in [22] to measure how much previous knowledge a model forgets after learning new tasks. They compared the forgetting ratio to average test error and demonstrated that the forgetting ratio could reflect the severity of the forgetting problem.…”

Section: Metricsmentioning

confidence: 99%

CLeaR: An Adaptive Continual Learning Framework for Regression Tasks

He,

Sick

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Catastrophic forgetting means that a trained neural network model gradually forgets the previously learned tasks when retrained on new tasks. Overcoming the forgetting problem is a major problem in machine learning. Numerous continual learning algorithms are very successful in incremental classification tasks, where new labels appear frequently. However, there is currently no research that addresses the catastrophic forgetting problem in regression tasks as far as we know. This problem has emerged as one of the primary constraints in some applications, such as renewable energy forecasts. This article clarifies the problem-related definitions and proposes a new methodological framework that can forecast regression task targets and update itself by continual learning. The framework consists of forecasting neural networks and buffers, which store newly collected data from a data stream in an application. Changes in the probability distribution of the data stream will be identified by the framework and learned sequentially. The framework is called CLeaR (Continual Learning for Regression Tasks), where components can be flexibly customized for a specific application scenario. We design two sets of experiments to evaluate the CLeaR framework concerning fitting error (training), prediction error (test), and forgetting ratio. The first one is based on an artificial time series to explore how hyperparameters affect the CLeaR framework. The second one is designed with data collected from European wind farms to evaluate the performance of the CLeaR framework in a real-world application. The experimental results demonstrate that the CLeaR framework can efficiently accumulate knowledge in the data stream and improve the prediction accuracy. The article concludes with further research issues arising from requirements to extend the framework.

show abstract

Visualizing the Behavior of Convolutional Neural Networks for Time Series Forecasting

Henze

Sick

2021

Studies in Computational Intelligence

View full text Add to dashboard Cite

Continuous Learning of Deep Neural Networks to Improve Forecasts for Regional Energy Markets

Cited by 11 publications

References 19 publications

Freight Rate and Demand Forecasting in Road Freight Transportation Using Econometric and Artificial Intelligence Methods

Freight Rate and Demand Forecasting in Road Freight Transportation Using Econometric and Artificial Intelligence Methods

CLeaR: An Adaptive Continual Learning Framework for Regression Tasks

Visualizing the Behavior of Convolutional Neural Networks for Time Series Forecasting

Contact Info

Product

Resources

About