Design of a bi-level PSO based modular neural network for multi-step time series prediction

被引:0
作者
Li, Wenjing [1 ,2 ,3 ,4 ]
Liu, Yonglei [1 ,2 ,3 ,4 ]
Chen, Zhiqian [1 ,2 ,3 ,4 ]
机构
[1] Beijing Univ Technol, Fac Informat Technol, Beijing 100124, Peoples R China
[2] Beijing Inst Artificial Intelligence, Beijing 100124, Peoples R China
[3] Beijing Key Lab Computat Intelligence & Intelligen, Beijing 100124, Peoples R China
[4] Beijing Lab Intelligent Environm Protect, Beijing 100124, Peoples R China
基金
中国国家自然科学基金;
关键词
Multi-step time series prediction; Modular neural network; Bi-level particle swarm optimization; Self-determined structure; Fuzzy rule;
D O I
10.1007/s10489-024-05638-0
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Derived from an effective strategy - direct and multiple-input multiple-output strategy, a modular neural network based on a bi-level particle swarm optimization algorithm (BLPSO-MNN) is proposed in the present study to improve the accuracy for multi-step time series prediction. While a binary particle swarm optimization algorithm is designed for the external layer to optimize the task division of prediction horizons, a multi-objective particle swarm optimization algorithm is designed for the internal layer to trade off between the prediction accuracy and structural complexity for each subnetwork in modular neural network. Besides, a set of fuzzy If-Then rules is proposed to determine the historical information to be input to subnetworks. Thus, the structure of BLPSO-MNN, including the number of modules as well as the subnetwork structure, is self-determined accordingly. Numerous experiments are conducted for 18-step-ahead time series prediction to evaluate the performance of BLPSO-MNN. Experimental results show that, although the prediction accuracy decreases when the prediction horizon is large, the overall performance of BLPSO-MNN is superior over all comparative models with greater improvement for larger horizons, indicating it is suitable for a long-term prediction. Besides, the set of fuzzy rules balances the prediction accuracy against the structural complexity caused by the subnetwork inputs.
引用
收藏
页码:8612 / 8633
页数:22
相关论文
共 47 条
[1]  
Antti S., 2006, BRUGES BELGIUM, V6, P143
[2]   PSO-MISMO Modeling Strategy for MultiStep-Ahead Time Series Prediction [J].
Bao, Yukun ;
Xiong, Tao ;
Hu, Zhongyi .
IEEE TRANSACTIONS ON CYBERNETICS, 2014, 44 (05) :655-668
[3]  
Ben Taieb Souhaib, 2009, Proceedings 2009 International Joint Conference on Neural Networks (IJCNN 2009 - Atlanta), P3054, DOI 10.1109/IJCNN.2009.5178802
[4]   Multiple-output modeling for multi-step-ahead time series forecasting [J].
Ben Taieb, Souhaib ;
Sorjamaa, Antti ;
Bontempi, Gianluca .
NEUROCOMPUTING, 2010, 73 (10-12) :1950-1957
[5]   Extending machine learning prediction capabilities by explainable AI in financial time series prediction [J].
Celik, Taha Bugra ;
Ican, Ozgur ;
Bulut, Elif .
APPLIED SOFT COMPUTING, 2023, 132
[6]   Research on Traffic Flow Prediction in the Big Data Environment Based on the Improved RBF Neural Network [J].
Chen, Dawei .
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, 2017, 13 (04) :2000-2008
[7]   A novel MOPSO-SODE algorithm for solving three-objective SR-ES-TR portfolio optimization problem [J].
Chen, Yinnan ;
Zhao, Xinchao ;
Hao, Junling .
EXPERT SYSTEMS WITH APPLICATIONS, 2023, 233
[8]   Temporal Attention-Augmented Bilinear Network for Financial Time-Series Data Analysis [J].
Dat Thanh Tran ;
Iosifidis, Alexandros ;
Kanniainen, Juho ;
Gabbouj, Moncef .
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, 2019, 30 (05) :1407-1418
[9]   Multi-scale adaptive attention-based time-variant neural networks for multi-step time series forecasting [J].
Gao, Changxia ;
Zhang, Ning ;
Li, Youru ;
Lin, Yan ;
Wan, Huaiyu .
APPLIED INTELLIGENCE, 2023, 53 (23) :28974-28993
[10]   Multi-step forecasting of multivariate time series using multi-attention collaborative network [J].
He, Xiaoyu ;
Shi, Suixiang ;
Geng, Xiulin ;
Yu, Jie ;
Xu, Lingyu .
EXPERT SYSTEMS WITH APPLICATIONS, 2023, 211