As a representative commodity of industrial commodities, steel price research can help the stable development of the steel industry. To explore the dynamics of steel prices, a new combined model named EEMD-TRANSFORMER-ARIMA is proposed, which is based on the Ensemble Empirical Mode Decomposition (EEMD) method and Transformer attention mechanism and Autoregressive Integrated Moving Average (ARIMA) model. EEMD decomposes the steel price time series data, the stability of the decomposed component data is tested, and the Transformer model and ARIMA model are used for prediction. The experiment indicates that the combined models take full advantage of the strong points of several single models and deal with non-stationary and non-linear time series. The result shows that EEMD-TRANSFORMER-ARIMA has excellent predicting performance for all six steel price data. It will be an effective method of prediction that will assist the government and enterprises in making market decisions.
XUX J, ZHANGY. Regional steel price index forecasts with neural networks: Evidence from east, south, north, central south, northeast, southwest, and Northwest China[J]. The Journal of Supercomputing, 2023, 79(12): 13601-13619. DOI: 10.1007/s11227-023-05207-1 .
ZHANGZ L, LIUM F, HUH J. Emotion prediction in english-chinese bilingual text based on joint feature and Bi-LSTM model[J]. Journal of Wuhan University (Natural Science Edition), 2019, 65(3): 269-275. DOI: 10.14188/j.1671-8836.2019.03.006(Ch ).
XUY, ZHUL Q, HUANGB, et al. Public opinion analysis based on EEMD-transformer model: Taking COVID-19 public opinion as an example[J]. Journal of Wuhan University (Natural Science Edition), 2020, 66(5): 418-424. DOI: 10.14188/j.1671-8836.2020.0147(Ch ).
[11]
HUANGN E, SHENZ, LONGS R, et al. The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis[J]. Proceedings of the Royal Society of London Series A: Mathematical, Physical and Engineering Sciences, 1998, 454(1971): 903-995. DOI: 10.1098/rspa.1998.0193 .
WANGW B, FEIP S, YIX M. Prediction of China stock market based on EMD and neural network[J]. Systems Engineering-Theory & Practice, 2010, 30(6): 1027-1033 (Ch). DOI: 10.12011/1000-6788(2010)6-1027 .
[14]
SHYUY W, CHANGC C. A hybrid model of MEMD and PSO-LSSVR for steel price forecasting[J]. International Journal of Engineering and Management Research, 2022, 12(1): 30-40. DOI: 10.31033/ijemr.12.1.5 .
[15]
WUZ H, HUANGN E. Ensemble empirical mode decomposition: A noise-assisted data analysis method[J]. Advances in Adaptive Data Analysis, 2009, 1(1): 1-41. DOI: 10.1142/s1793536909000047 .
ZHOUX G, HEQ, HUANGX X. Portfolio selection model and empirical study based on EEMD and fuzzy regression[J]. Operations Research and Management Science, 2022, 31(5): 150-155. DOI: 10.12005/orms.2022.0162(Ch ).
YANGJ L, TANGG Q, ZHANGJ W. Research on stock price prediction of China and America based on EEMD-elman-adaboost[J]. Operations Research and Management Science, 2022, 31(11): 194-199. DOI: 10.12005/orms.2022.0373(Ch ).
[20]
VASWANIA, SHAZEERN, PARMARN, et al. Attention is all you need[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. New York: ACM, 2017: 6000-6010. DOI: 10.5555/3295222.3295349 .
[21]
BOXG E P, PIERCED A. Distribution of residual autocorrelations in autoregressive-integrated moving average time series models[J]. Journal of the American Statistical Association, 1970, 65(332): 1509-1526. DOI: 10.1080/01621459.1970.10481180 .
[22]
WUH X, XUJ H, WANGJ M, et al. Autoformer: Decomposition Transformers with Auto-correlation for Long-term Series Forecasting[EB/OL]. [2021-12-31].