Ashraf, Waqar Muhammad;
Dua, Vivek;
(2024)
Storage of weights and retrieval method (SWARM) approach for neural networks hybridized with conformal prediction to construct the prediction intervals for energy system applications.
International Journal of Data Science and Analytics
10.1007/s41060-024-00595-w.
(In press).
Preview |
Text
Ashraf_s41060-024-00595-w.pdf Download (2MB) | Preview |
Abstract
The prediction intervals represent the uncertainty associated with the model-predicted responses that impacts the sequential decision-making analytics. Here in this work, we present a novel model-based data-driven approach to construct the prediction intervals around the model-simulated responses using artificial neural network (ANN) model. The loss function is modified with least mean square error and standard deviation between the model-simulated and actual responses for the online-training mode of ANN model development. The parameters (weights and biases) stored during the model development are extracted and are deployed to construct the prediction intervals with 95% confidence level for the test datasets of the three energy systems-based case studies including: crease recovery angle, energy efficiency cooling & energy efficiency heating and gas turbine power plant & coal power plant which are taken from literature, benchmark datasets and industrial-scale applications, respectively. The developed ANN models present root-mean-squared error of 1.20% and 0.52% on test dataset for energy efficiency cooling and energy efficiency heating, respectively. The width of prediction intervals made by the proposed approach, called as Storage of Weights And Retrieval Method (SWARM), incorporates the information available for each test observation during the model training and the SWARM-based prediction intervals are compared to those of inductive conformal prediction (ICP) technique. It is noted that SWARM technique offers better locally adaptive prediction intervals than those of ICP, highlighting the effectiveness of the SWARM technique for the estimation of prediction intervals for the case studies. This research presents a novel data-driven approach to construct the prediction intervals using the model-based information that can be applied on different real-life applications.
| Type: | Article |
|---|---|
| Title: | Storage of weights and retrieval method (SWARM) approach for neural networks hybridized with conformal prediction to construct the prediction intervals for energy system applications |
| Open access status: | An open access version is available from UCL Discovery |
| DOI: | 10.1007/s41060-024-00595-w |
| Publisher version: | http://dx.doi.org/10.1007/s41060-024-00595-w |
| Language: | English |
| Additional information: | This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. |
| Keywords: | Prediction interval; Uncertainty quantification; Conformal prediction; SWARM |
| UCL classification: | UCL UCL > Provost and Vice Provost Offices > UCL BEAMS UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Engineering Science > Dept of Chemical Engineering |
| URI: | https://discovery.ucl.ac.uk/id/eprint/10196897 |
Archive Staff Only
 |
View Item |
