EMD-Based Recurrent Neural Network with Adaptive Regrouping for Port Cargo Throughput Prediction
Accurate prediction of port cargo throughput (PCT) plays an important role in economic investment, transportation planning, port planning and design, etc. PCT time series have the properties of nonlinearity and complexity. To guarantee high-quality prediction performance, we propose to first adopt the empirical mode decomposition (EMD) to decompose the original PCT time series into high and low frequency components. It is more difficult to predict some components due to their properties of weak mathematical regularity. To take advantage of the selfsimilarities within components, each component will be divided into several small parts which are adaptive regrouped (ARG) via the standardized euclidean distance (SED)-based similarity measure. The regrouped parts are then selected to form the training dataset for long short-term memory (LSTM) to enhance the prediction accuracy of each component. The final prediction result can be obtained by integrating the predicted components. Our proposed three-step prediction framework (called EMD-ARG-LSTM) benefits from the property decomposition and adaptive similarity regrouping. Experimental results have illustrated the superior performance of the proposed method in terms of both prediction accuracy and robustness.
KeywordsPort cargo throughput Prediction Long short-term memory Empirical mode decomposition Similarity regrouping
This work was supported by National Natural Science Foundation of China (Nos.: 51609195 and 51479156), Fund of Hubei Key Laboratory of Transportation Internet of Things (No.: WHUTIOT-2017B003), and Independent Innovation Research Funding for Undergraduates (No.: 2018-HY-A1-01).
- 1.Zhang, C., Huang, L., Zhao, Z.: Research on combination forecast of port cargo throughput based on time series and causality analysis. J. Ind. Eng. Manag. 6(1), 124–134 (2013)Google Scholar
- 9.Liu, R.W., Chen, J., Liu, Z., Li, Y., Liu, Y., Liu, J.: Vessel traffic flow separation-prediction using low-rank and sparse decomposition. In: IEEE ITSC, pp. 1–6 (2017)Google Scholar
- 10.Odom, M.D., Sharda, R.: Stock market prediction system with modular neural networks. In: IEEE IJCNN, pp. 1–6 (1990)Google Scholar
- 12.Lv, Y., Duan, Y., Kang, W., Li, Z., Wang, F.Y.: Traffic flow prediction with big data: a deep learning approach. IEEE Trans. Intell. Transp. Syst. 16(2), 865–873 (2015)Google Scholar