融合MIC与Res-LSTM模型的有效波高预测

doi:10.11978/2023113

Abstract

Abstract:

The prediction of significant wave height (SWH) plays an important role in marine transportation and maritime activities. Based on the near-shore real measurement data of the Hailing Island, Yangjiang, China, a network model integrating the maximum information coefficient algorithm (MIC), residual network (ResNet) and long and short-term memory network (LSTM) is proposed. Firstly, the MIC algorithm was used to screen out the parameters with high correlation with the target predictors from the dataset as the input of the model. Then the residual network was introduced into the LSTM to construct the Res-LSTM prediction model. Finally, the r-squared (R²), root mean square error (RMSE), mean absolute error (MAE) and mean absolute percentage error (MAPE) were selected to evaluate the prediction results. Meanwhile, the prediction results of extreme gradient boosting (XGBoost) network, support vector regression (SVR) network and LSTM network were compared. The results demonstrate that the MIC-Res-LSTM model can improve the accuracy of the short-time significant wave height prediction values.

Key words: wave height prediction, maximum information coefficient, residual network, long and short-term memory network, support vector regression

ZHU Daoheng, LI Yan, LI Zhiqiang, LIU Run. Fusing MIC and Res-LSTM models for significant wave height prediction[J].Journal of Tropical Oceanography, 2024, 43(4): 76-85.

Figures/Tables 11

Tab. 1

Fig. 1

Fig. 2

Fig. 3

Tab. 2

Tab. 3

Tab. 4

Fig. 4

Tab. 5

Tab. 6

Fig. 5

References 30

[1]	周水华, 洪晓, 梁昌霞, 等, 2020. 基于人工神经网络的台风浪高快速计算方法[J]. 热带海洋学报, 39(4): 25-33. doi: 10.11978/2019089
	ZHOU SHUIHUA, HONG XIAO, LIANG CHANGXIA, et al, 2020. A method of tropical cyclone wave height calculation based on Artificial Neural Network[J]. Journal of Tropical Oceanography, 39(4): 25-33 (in Chinese with English abstract). doi: 10.11978/2019089
[2]	朱道恒, 李志强, 2021. 最大互信息系数的并行计算方法研究[J]. 科学技术与工程, 21(34): 14625-14633.
	ZHU DAOHENG, LI ZHIQIANG, 2021. Parallel calculation method for maximum information coefficient[J]. Science Technology and Engineering, 21(34): 14625-14633 (in Chinese with English abstract).
[3]	ABHIGNA P, JERRITTA S, SRINIVASAN R, et al, 2017. Analysis of feed forward and recurrent neural networks in predicting the significant wave height at the moored buoys in Bay of Bengal[C]// Proceedings of the 2017 international conference on communication and signal processing (ICCSP). Chennai: IEEE: 1856-1860.
[4]	BENGIO Y, SIMARD P, FRASCONI P, 1994. Learning long-term dependencies with gradient descent is difficult[J]. IEEE Transactions on Neural Networks, 5(2): 157-166. doi: 10.1109/72.279181 pmid: 18267787
[5]	CORNEJO-BUENO L, NIETO-BORGE J C, GARCÍA-DÍAZ P, et al, 2016. Significant wave height and energy flux prediction for marine energy applications: a grouping genetic algorithm—extreme learning machine approach[J]. Renewable Energy, 97: 380-389.
[6]	DEKA P C, PRAHLADA R, 2012. Discrete wavelet neural network approach in significant wave height forecasting for multistep lead time[J]. Ocean Engineering, 43: 32-42.
[7]	FAN SHUNTAO, XIAO NIANHAO, DONG SHENG, 2020. A novel model to predict significant wave height based on long short-term memory network[J]. Ocean Engineering, 205: 107298.
[8]	FENG ZHIJIE, HU PO, LI SHUIQING, et al, 2022. Prediction of significant wave height in offshore China based on the machine learning method[J]. Journal of Marine Science and Engineering, 10(6): 836.
[9]	GAO SONG, ZHAO PENG, PAN BIN, et al, 2018. A nowcasting model for the prediction of typhoon tracks based on a long short term memory neural network[J]. Acta Oceanologica Sinica, 37(5): 8-12.
[10]	GAO SONG, HUANG JUAN, LI YARU, et al, 2021. A forecasting model for wave heights based on a long short-term memory neural network[J]. Acta Oceanologica Sinica, 40(1): 62-69.
[11]	GÜNER H A A, YÜKSEL Y, ÇEVIK E Ö, 2013. Estimation of wave parameters based on nearshore wind-wave correlations[J]. Ocean Engineering, 63: 52-62.
[12]	HASHIM R, ROY C, MOTAMEDI S, et al, 2016. Selection of climatic parameters affecting wave height prediction using an enhanced Takagi-Sugeno-based fuzzy methodology[J]. Renewable and Sustainable Energy Reviews, 60: 246-257.
[13]	HE KAIMING, ZHANG XIANGYU, REN SHAOQING, et al, 2016. Deep residual learning for image recognition[C]// IEEE conference on computer vision and pattern recognition. Las Vegas: IEEE: 770-778.
[14]	HOCHREITER S, SCHMIDHUBER J, 1997. Long short-term memory[J]. Neural Computation, 9(8): 1735-1780. doi: 10.1162/neco.1997.9.8.1735 pmid: 9377276
[15]	LIU HUI, MI XIWEI, LI YANFEI, et al, 2019. Smart wind speed deep learning based multi-step forecasting model using singular spectrum analysis, convolutional Gated Recurrent Unit network and Support Vector Regression[J]. Renewable Energy, 143: 842-854.
[16]	MA KEZHAO, KONG JIA, WANG YIHAN, et al, 2022. Review of the applications of Kalman filtering in quantum systems[J]. Symmetry, 14(12): 2478.
[17]	MAHJOOBI J, ETEMAD-SHAHIDI A, KAZEMINEZHAD M H, 2008. Hindcasting of wave parameters using different soft computing methods[J]. Applied Ocean Research, 30(1): 28-36.
[18]	MAHJOOBI J, ADELI MOSABBEB E, 2009. Prediction of significant wave height using regressive support vector machines[J]. Ocean Engineering, 36(5): 339-347.
[19]	MAKARYNSKYY O, PIRES-SILVA A A, MAKARYNSKA D, et al, 2005. Artificial neural networks in wave predictions at the west coast of Portugal[J]. Computers & Geosciences, 31(4): 415-424.
[20]	ÖZGER M, 2010. Significant wave height forecasting using wavelet fuzzy logic approach[J]. Ocean Engineering, 37(16): 1443-1451.
[21]	RESHEF D N, RESHEF Y A, FINUCANE H K, et al, 2011. Detecting novel associations in large data sets[J]. Science, 334(6062): 1518-1524. doi: 10.1126/science.1205438 pmid: 22174245
[22]	RIZAL A M, NINGSIH N S, 2020. Ocean wave energy potential along the west coast of the Sumatra island, Indonesia[J]. Journal of Ocean Engineering and Marine Energy, 6(2): 137-154.
[23]	RODRIGUEZ A, SÁNCHEZ-ARCILLA A, REDONDO J M, et al, 1995. Pollutant dispersion in the nearshore region: modelling and measurements[J]. Water Science and Technology, 32(9-10): 169-178.
[24]	SHAMSHIRBAND S, MOSAVI A, RABCZUK T, et al, 2020. Prediction of significant wave height; comparison between nested grid numerical model, and machine learning models of artificial neural networks, extreme learning and support vector machines[J]. Engineering Applications of Computational Fluid Mechanics, 14(1): 805-817.
[25]	STOKES G C, 2010. On the theory of oscillation wave[M/OL]. STOKES G C. Mathematical and Physical Papers. Cambridge: Cambridge University Press: 197-229. https://doi.org/10.1017/CBO9780511702242.
[26]	THEODOROPOULOS P, SPANDONIDIS C C, THEMELIS N, et al, 2021. Evaluation of different deep-learning models for the prediction of a ship’s propulsion power[J]. Journal of Marine Science and Engineering, 9(2): 116.
[27]	WANG JICHAO, WANG YUE, YANG JUNGANG, 2021. Forecasting of significant wave height based on gated recurrent unit network in the Taiwan strait and its adjacent waters[J]. Water, 13(1): 86.
[28]	YANG SHAOBO, ZHANG ZHENQUAN, FAN LINLIN, et al, 2019. Long-term prediction of significant wave height based on SARIMA model in the South China Sea and adjacent waters[J]. IEEE Access, 7: 88082-88092.
[29]	YANG SHILUN, SHI BENWEI, BOUMA T J, et al, 2012. Wave attenuation at a salt marsh margin: a case study of an exposed coast on the Yangtze Estuary[J]. Estuaries and Coasts, 35(1): 169-182.
[30]	ZHANG DA, KABUKA M R, 2018. Combining weather condition data to predict traffic flow: a GRU-based deep learning approach[J]. IET Intelligent Transport Systems, 12(7): 578-585.

指标	大气温度/℃	风向/°	风速/(m·s^-1)	大气压强/kpa	相对湿度/%	水温/℃	潮高/m	深水波高/m	深水波周期/s	水深/m	平均波周期/s	有效波高/m
平均值	30.23	200.62	1.31	99.89	76.24	28.79	1.68	0.71	4.68	0.93	4.95	0.32
最大值	35.00	359.00	4.30	100.19	88.10	36.92	3.32	1.03	6.41	2.54	9.76	0.92
最小值	26.00	0	0	99.59	35.40	25.00	0.25	0.39	3.53	0.21	2.46	0.08
标准偏差	1.98	93.99	0.84	1.22	7.17	1.98	66.87	0.15	0.63	0.53	0.95	0.16
变异系数/%	6.54	46.85	64.31	0.12	9.41	6.90	39.80	20.73	13.43	57.26	19.19	51.13

参数	batch_size	learning_rate	epoch	beta1	beta2	epsilon	timesteps	dropout
自测数据集实验	72	0.0025	100	0.9	0.999	1×10⁻⁸	12	0.001
公开数据集实验	128	0.001	150	0.9	0.999	1×10⁻⁸	12	0.001

特征指标	MIC	Pearson	Spearman	Kendall
大气温度	0.176	0.026	0.001	0.002
风向	0.154	0.003	0.015	0.010
风速	0.177	0.010	0.047	0.031
大气压强	0.202	0.003	0.108	0.072
相对湿度	0.173	0.019	0.031	0.002
水温	0.221	0.361	0.477	0.324
潮高	0.487	0.647	0.677	0.506
深水波高	0.379	0.286	0.293	0.199
深水波周期	0.361	0.432	0.401	0.349
平均波周期	0.473	0.298	0.312	0.217
水深	0.659	0.672	0.521	0.497

模型	时间跨度	R²	RMSE	MAE	MAPE/%
XGboost	10min	0.7761	0.1084	0.0903	38.47
SVR		0.8256	0.1152	0.1017	45.59
LSTM		0.9653	0.0373	0.0293	13.35
Res-LSTM		0.9814	0.0194	0.0152	6.09
XGboost	30min	0.7345	0.1473	0.1049	32.16
SVR		0.8136	0.1137	0.0925	37.84
LSTM		0.9567	0.0812	0.0675	14.60
Res-LSTM		0.9641	0.0385	0.0141	6.79
XGboost	1h	0.7198	0.1266	0.1044	43.54
SVR		0.7761	0.0701	0.0572	23.72
LSTM		0.9138	0.0575	0.0524	23.00
Res-LSTM		0.9423	0.0516	0.0416	17.70

变量名	MIC	Pearson	Spearman	Kendall
月份	0.1560	0.1011	0.0079	0.0033
小时	0.0744	0.0219	0.0690	0.0471
分钟	0.0663	0.0059	0.0071	0.0047
风向	0	0	0	0
平均风速	0.1847	0.0578	0.0426	0.0319
峰值风速	0.4814	0.5973	0.6078	0.4462
主波周期	0.3148	0.6088	0.6247	0.4572
平均波周期	0.1801	0.0980	0.0248	0.0088
主波方向	0.2276	0.5017	0.3329	0.2420
海平面压力	0.1477	0.1555	0.0900	0.0566
空气温度	0.0975	0.2818	0.1524	0.0989
海平面温度	0.1605	0.2554	0.3597	0.2030
露点温度	0.1864	0.2079	0.2978	0.2030
浮标能见度	0.1617	0.0353	0.3190	0.2146

Fusing MIC and Res-LSTM models for significant wave height prediction

RichHTML

PDF (PC)

Like

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 11

References 30

Related Articles 0

Metrics

Comments

Recommended 0

模型	相关性阈值	R²	RMSE	MAE	MAPE/%
XGBoost	MIC=0	0.8271	0.4635	0.2711	22.22
	MIC=0.2	0.8483	0.5129	0.2889	21.15
	MIC=0.3	0.8596	0.5265	0.3323	20.39
SVR	MIC=0	0.7686	0.5069	0.3854	40.89
	MIC=0.2	0.8108	0.5601	0.3555	32.84
	MIC=0.3	0.8837	0.6092	0.3507	25.56
LSTM	MIC=0	0.9819	0.1733	0.1312	14.24
	MIC=0.2	0.9865	0.1346	0.0933	9.44
	MIC=0.3	0.9879	0.1163	0.0771	7.31
Res-LSTM	MIC=0	0.9817	0.1349	0.0945	10.60
	MIC=0.2	0.9916	0.1089	0.0652	8.50
	MIC=0.3	0.9932	0.0955	0.0644	6.48