Premium
A hybrid model for financial time‐series forecasting based on mixed methodologies
Author(s) -
Luo Zhidan,
Guo Wei,
Liu Qingfu,
Zhang Zhengjun
Publication year - 2021
Publication title -
expert systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.365
H-Index - 38
eISSN - 1468-0394
pISSN - 0266-4720
DOI - 10.1111/exsy.12633
Subject(s) - autoregressive integrated moving average , computer science , benchmark (surveying) , series (stratigraphy) , nonlinear system , autoregressive model , time series , hilbert–huang transform , taylor series , finance , econometrics , mathematics , machine learning , economics , paleontology , mathematical analysis , physics , geodesy , filter (signal processing) , quantum mechanics , computer vision , biology , geography
This paper proposes a hybrid model that combines ensemble empirical mode decomposition (EEMD), autoregressive integrated moving average (ARIMA), and Taylor expansion using a tracking differentiator to forecast financial time series. Specifically, the financial time series is decomposed by EEMD into some subseries. Then, the linear portion of each subseries is forecasted by the linear ARIMA model, while the nonlinear portion is predicted by the nonlinear Taylor expansion model. The forecasting results of the linear and nonlinear models are combined into the predicted result of each subseries. The final prediction result is obtained by combining the prediction values of all the subseries. The empirical results with real financial time series data demonstrate that this new hybrid approach outperforms the benchmark hybrid models considered in this paper.