A perturbative approach for enhancing the performance of time series forecasting

被引:11
作者
de Mattos Neto, Paulo S. G. [1 ]
Ferreira, Tiago A. E. [2 ]
Lima, Aranildo R. [3 ]
Vasconcelos, Germano C. [1 ]
Cavalcanti, George D. C. [1 ]
机构
[1] Univ Fed Pernambuco, Ctr Informat, Recife, PE, Brazil
[2] Univ Fed Rural Pernambuco UFRPE, Dept Estat & Informat, Recife, PE, Brazil
[3] Univ British Columbia, Dept Earth Ocean & Atmospher Sci, Vancouver, BC, Canada
关键词
Perturbation theory; Time series forecasting; Combination of forecasts; Artificial neural networks; NEURAL-NETWORKS; HYBRID ARIMA; MODEL; PREDICTION; ALGORITHM; METHODOLOGY; PARAMETERS; SELECTION; FUSION;
D O I
10.1016/j.neunet.2017.02.004
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
This paper proposes a method to perform time series prediction based on perturbation theory. The approach is based on continuously adjusting an initial forecasting model to asymptotically approximate a desired time series model. First, a predictive model generates an initial forecasting for a time series. Second, a residual time series is calculated as the difference between the original time series and the initial forecasting. If that residual series is not white noise, then it can be used to improve the accuracy of the initial model and a new predictive model is adjusted using residual series. The whole process is repeated until convergence or the residual series becomes white noise. The output of the method is then given by summing up the outputs of all trained predictive models in a perturbative sense. To test the method, an experimental investigation was conducted on six real world time series. A comparison was made with six other methods experimented and ten other results found in the literature. Results show that not only the performance of the initial model is significantly improved but also the proposed method outperforms the other results previously published. (C) 2017 Elsevier Ltd. All rights reserved.
引用
收藏
页码:114 / 124
页数:11
相关论文
共 50 条
[41]   Improving forecasting by subsampling seasonal time series [J].
Li, Xixi ;
Petropoulos, Fotios ;
Kang, Yanfei .
INTERNATIONAL JOURNAL OF PRODUCTION RESEARCH, 2023, 61 (03) :976-992
[42]   Forecasting nonlinear time series with a hybrid methodology [J].
Aladag, Cagdas Hakan ;
Egrioglu, Erol ;
Kadilar, Cem .
APPLIED MATHEMATICS LETTERS, 2009, 22 (09) :1467-1470
[43]   Deep Video Prediction for Time Series Forecasting [J].
Zeng, Zhen ;
Balch, Tucker ;
Veloso, Manuela .
ICAIF 2021: THE SECOND ACM INTERNATIONAL CONFERENCE ON AI IN FINANCE, 2021,
[44]   Modeling ANNs Performance on Time Series Forecasting [J].
Suarez, Ranyart R. ;
Graff, Mario .
2013 IEEE INTERNATIONAL AUTUMN MEETING ON POWER, ELECTRONICS AND COMPUTING (ROPEC), 2013,
[45]   Intuitionistic Fuzzy Time Series: An Approach for Handling Nondeterminism in Time Series Forecasting [J].
Kumar, Sanjay ;
Gangwar, Sukhdev Singh .
IEEE TRANSACTIONS ON FUZZY SYSTEMS, 2016, 24 (06) :1270-1281
[46]   Time series analysis and short-term forecasting of solar irradiation, a new hybrid approach [J].
Bigdeli, N. ;
Borujeni, M. Salehi ;
Afshar, K. .
SWARM AND EVOLUTIONARY COMPUTATION, 2017, 34 :75-88
[47]   A Time Series Forecasting Approach Based on Nonlinear Spiking Neural Systems [J].
Long, Lifan ;
Liu, Qian ;
Peng, Hong ;
Yang, Qian ;
Luo, Xiaohui ;
Wang, Jun ;
Song, Xiaoxiao .
INTERNATIONAL JOURNAL OF NEURAL SYSTEMS, 2022, 32 (08)
[48]   Nonlinear time-series forecasting: A fuzzy-neural approach [J].
Nie, JH .
NEUROCOMPUTING, 1997, 16 (01) :63-76
[49]   A Model Ranking Based Selective Ensemble Approach for Time Series Forecasting [J].
Adhikari, Ratnadip ;
Verma, Ghanshyam ;
Khandelwal, Ina .
INTERNATIONAL CONFERENCE ON COMPUTER, COMMUNICATION AND CONVERGENCE (ICCC 2015), 2015, 48 :14-21
[50]   Enhancing financial time series forecasting in the shipping market: A hybrid approach with Light Gradient Boosting Machine [J].
Song, Xuefei ;
Chen, Zhong Shuo .
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE, 2024, 136