A hybrid application of soft computing methods with wavelet SVM and neural network to electric power load forecasting

Machine learning methods such as Support Vector Machine (SVM) and Neural Network (NN) as soft computing methods are widely used to solve nonlinear problems. Wavelet analysis and artificial intelligence machine learning will be combined to improve the self learning ability and prediction accuracy. Ac...

Full description

Saved in:
Bibliographic Details
Published in:Journal of Electrical Systems and Information Technology 2018-12, Vol.5 (3), p.681-696
Main Authors: Xia, Changhao, Zhang, Mi, Cao, Jin
Format: Article
Language:English
Subjects:
Load forecasting
Neural network
Soft computing
SVM
Wavelet analysis
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Machine learning methods such as Support Vector Machine (SVM) and Neural Network (NN) as soft computing methods are widely used to solve nonlinear problems. Wavelet analysis and artificial intelligence machine learning will be combined to improve the self learning ability and prediction accuracy. Actual historical load data is decomposed into high and low frequency load sequence by using wavelet analysis. Utilizing SVM and NN machine learning methods, choosing the appropriate parameters such as network structures, penalty parameter and kernel function width by optimization program, the single branch predictions for each sequence are separately made, and each branch prediction results are reconstructed to achieve ultimate load forecasting. Application result shows that wavelet SVM has higher prediction accuracy.
ISSN:2314-7172
2314-7172
DOI:10.1016/j.jesit.2017.05.008