Fuzzy-based weighting long short-term memory network for demand forecasting

One of the main challenges in short-term electrical load forecasting is extraction of nonlinear relationships and complex dependencies among different time instances of the load time series. To deal with this difficulty, a hybrid forecasting method is proposed in this paper that uses the fuzzy exper...

Full description

Saved in:
Bibliographic Details
Published in:The Journal of supercomputing Vol. 79; no. 1; pp. 435 - 460
Main Author: Imani, Maryam
Format: Journal Article
Language:English
Published: New York Springer US 01.01.2023
Springer Nature B.V
Subjects:
Artificial intelligence
Compilers
Computer Science
Deep learning
Electrical loads
Expert systems
Forecasting
Genetic algorithms
Interpreters
Mathematics
Methods
Neural networks
Power
Processor Architectures
Programming Languages
Time series
Wavelet transforms
Fuzzy logic
LSTM
Load forecasting
Weather conditions
ISSN:0920-8542, 1573-0484
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:One of the main challenges in short-term electrical load forecasting is extraction of nonlinear relationships and complex dependencies among different time instances of the load time series. To deal with this difficulty, a hybrid forecasting method is proposed in this paper that uses the fuzzy expert systems and deep learning methods. In the first step, dependency of previous time instances to the next instance to be load forecasted is achieved through a fuzzy system with 125 rules. Then, the obtained weights are used beside the actual load values as the input of a long short-term memory network for load forecasting. The obtained results on two popular datasets show the superior performance of the proposed method in terms of various evaluation measures.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0920-8542
1573-0484
DOI:10.1007/s11227-022-04659-1