Application of neural networks and support vector machine for significant wave height prediction

Abstract. Accurate and significant wave height prediction with a couple of hours of warning time should offer major safety improvements for coastal and ocean engineering applications. However, significant wave height phenomenon is nonlinear and nonstationary, which makes any prediction simulation a non straightforward task. The aim of the research presented in this paper is to improve predicted significant wave height via a hybrid algorithm. Firstly, empirical mode decomposition (EMD) is used to preprocess the nonlinear data, which are decomposed into several simple signals. Then, least square support vector machine (LSSVM) with nonlinear learning ability is used to predict the significant wave height, and particle swarm optimization (PSO) is implemented to automatically perform the parameter selection in LSSVM modeling. The EMD-PSO-LSSVM model is used to predict the significant wave height for 1, 3 and 6 hours leading times of two stations in the offshore and deep-sea areas of the North Atlantic Ocean. The results show that the EMD-PSO-LSSVM model can remove the lag in the prediction timing of the single prediction models. Furthermore, the prediction accuracy of the EMD-LSSVM model that has not been optimized in the deep-sea area has been greatly improved, an improvement of the prediction accuracy of Coefficient of determination (R2) from 0.991, 0.982 and 0.959 to 0.993, 0.987 and 0.965, respectively, has been observed. The proposed new hybrid model shows good accuracy and provides an effective way to predict the significant wave height for the deep-sea area.

Download Full-text

Significant Wave Height Prediction Based on MSFD Neural Network

2019 Tenth International Conference on Intelligent Control and Information Processing (ICICIP) ◽

10.1109/icicip47338.2019.9012194 ◽

2019 ◽

Author(s):

Huan Wang ◽

Dongyang Fu ◽

Shan Liao ◽

Guancheng Wang ◽

Xiuchun Xiao

Keyword(s):

Neural Network ◽

Wave Height ◽

Significant Wave Height ◽

Significant Wave ◽

Height Prediction

Download Full-text

Neural Networks Based Simulation of Significant Wave Height

Volume 4: Materials Technology; Ocean Engineering ◽

10.1115/omae2007-29171 ◽

2007 ◽

Cited By ~ 3

Author(s):

H. Bazargan ◽

H. Bahai ◽

A. Aminzadeh-Gohari ◽

A. Bazargan

Keyword(s):

Neural Networks ◽

Wave Height ◽

Significant Wave Height ◽

Optimization Technique ◽

Conditional Probability Density ◽

Significant Wave ◽

The Pacific ◽

Wave Heights ◽

Significant Wave Heights ◽

Long Term Forecasting

A large number of ocean activities call for real time or on-line forecasting of wind wave characteristics including significant wave height (Hs). The work reported in this paper uses statistics, and artificial neural networks trained with an optimization technique called simulated annealing to estimate the parameters of a probability distribution called hepta-parameter spline for the conditional probability density functions (pdf’s) of significant wave heights given their eight immediate preceding 3-hourly observed Hs’s. These pdf’s are used in the simulation of significant wave heights related to a location in the Pacific. The paper also deals with short and long term forecasting of Hs for the region through generating random variates from the spline distribution.

Download Full-text

Gap-Filling and Predicting Wave Parameters Using Support Vector Regression Method

Volume 6: Ocean Engineering ◽

10.1115/omae2011-49814 ◽

2011 ◽

Author(s):

Maziar Golestani ◽

Mostafa Zeinoddini

Keyword(s):

Support Vector Regression ◽

Wave Height ◽

Significant Wave Height ◽

Rational Design ◽

Wave Spectrum ◽

Support Vector ◽

Wave Spectra ◽

Significant Wave ◽

Wave Parameters ◽

Data Gaps

Knowledge of relevant oceanographic parameters is of utmost importance in the rational design of coastal structures and ports. Therefore, an accurate prediction of wave parameters is especially important for safety and economic reasons. Recently, statistical learning methods, such as Support Vector Regression (SVR) have been successfully employed by researchers in problems such as lake water level predictions, and significant wave height prediction. The current study reports potential application of a SVR approach to predict the wave spectra and significant wave height. Also the capability of the model to fill data gaps was tested using different approaches. Concurrent wind and wave records (standard meteorological and spectral density data) from 4 stations in 2003, 2007, 2008 and 2009 were used both for the training the SVR system and its verification. The choice of these four locations facilitated the comparison of model performances in different geographical areas. The SVR model was then used to obtain predictions for the wave spectra and also time series of wave parameters (separately for each station) such as its Hs and Tp from spectra and wind records. New approach was used to predict wave spectra comparing to similar studies. Reasonably well correlation was found between the predicted and measured wave parameters. The SVR model was first trained and tested using various methods for selecting training data. Also different values for SVM parameters (e.g. tolerance of termination criterion, cost, and gamma in kernel function) were tested. The best possible results were obtained using a Unix shell script (in Linux) which automatically implements different values for different input parameters and finds the best regression by calculating statistical scores like correlation of coefficient, RMSE, bias and scatter index. Finally for a better understanding of the results, Quantile-Quantile plots were produced. The results show that SVR can be successfully used for prediction of Hs and wave spectrum out of a series of wind and spectral wave parameters inputs. Also it was noticed that SVR is an efficient tool to be used when data gaps are present in the data.

Download Full-text

Sverdruv Munk Bretschneider Modification (SMB) for Significant Wave Height Prediction in Java Sea

British Journal of Applied Science & Technology ◽

10.9734/bjast/2016/19669 ◽

2016 ◽

Vol 16 (2) ◽

pp. 1-8 ◽

Cited By ~ 1

Author(s):

Aulia Aisjah ◽

Syamsul Arifin ◽

Wimala Danistha

Keyword(s):

Wave Height ◽

Significant Wave Height ◽

Significant Wave ◽

Height Prediction ◽

Java Sea

Download Full-text

Forecasting of Significant Wave Height Using Support Vector Regression

2015 Fifth International Conference on Advances in Computing and Communications (ICACC) ◽

10.1109/icacc.2015.109 ◽

2015 ◽

Author(s):

K. Ajeesh ◽

Paresh Chandra Deka

Keyword(s):

Support Vector Regression ◽

Wave Height ◽

Significant Wave Height ◽

Support Vector ◽

Significant Wave

Download Full-text

Significant Wave Height Prediction with the CRBM-DBN Model

Journal of Atmospheric and Oceanic Technology ◽

10.1175/jtech-d-18-0141.1 ◽

2019 ◽

Vol 36 (3) ◽

pp. 333-351 ◽

Cited By ~ 1

Author(s):

Xining Zhang ◽

Hao Dai

Keyword(s):

Wave Height ◽

Significant Wave Height ◽

Restricted Boltzmann Machine ◽

Series Data ◽

Model Parameters ◽

Boltzmann Machine ◽

Short Term ◽

Prediction Ability ◽

Significant Wave ◽

Height Prediction

AbstractIn recent years, deep learning technology has been gradually used for time series data prediction in various fields. In this paper, the restricted Boltzmann machine (RBM) in the classical deep belief network (DBN) is substituted with the conditional restricted Boltzmann machine (CRBM) containing temporal information, and the CRBM-DBN model is constructed. Key model parameters, which are determined by the particle swarm optimization (PSO) algorithm, are used to predict the significant wave height. Observed data in 2016, which are from nearshore and offshore buoys (i.e., 42020 and 42001) belonging to the National Data Buoy Center (NDBC), are taken to train the model, and the corresponding data in 2017 are used for testing with lead times of 1–24 h. In addition, we trained the data of 42040 in 2003 and tested the data in 2004 in order to investigate the prediction ability of the CRBM-DBN model for the extreme event. The prediction ability of the model is evaluated by the Nash–Sutcliffe coefficient of efficiency (CE) and root-mean-square error (RMSE). Experiments demonstrate that for the short-term (≤9 h) prediction, the RMSE and CE for the significant wave height prediction are <10 cm and >0.98, respectively. Moreover, the relative error of the short-term prediction for the maximum wave height is less than 26%. The excellent short-term and extreme events forecasting ability of the CRBM-DBN model is vital to ocean engineering applications, especially for designs of ocean structures and vessels.

Download Full-text

Prediction of the wave power in the Black Sea based on wind speed using artificial neural networks

E3S Web of Conferences ◽

10.1051/e3sconf/20185101006 ◽

2018 ◽

Vol 51 ◽

pp. 01006

Author(s):

Sorin Ciortan ◽

Eugen Rusu

Keyword(s):

Neural Networks ◽

Artificial Neural Networks ◽

Wind Speed ◽

Wave Height ◽

Black Sea ◽

Significant Wave Height ◽

The Black Sea ◽

Wave Power ◽

Significant Wave ◽

Artificial Neural

The paper proposes a prediction methodology for the significant wave height (and implicitly the wave power), based on the artificial neural networks. The proposed approach takes as input data the wind speed values recorded for different time periods. The prediction of significant wave height is useful both for assessment of wave energy as also for marine equipment design and navigation. The data used cover the time interval 1999 to 2007 and it was measured on Gloria drilling unit, which operates in the Romanian nearshore of the Black Sea at about 500 meters depth.

Download Full-text

Correlation Enhanced Machine Learning Approach based Wave Height Prediction

SMART MOVES JOURNAL IJOSCIENCE ◽

10.24113/ijoscience.v4i5.136 ◽

2018 ◽

Vol 4 (5) ◽

pp. 10

Author(s):

Ruchi Shrivastava ◽

Dr. Krishna Teerth Chaturvedi

Keyword(s):

Neural Network ◽

Machine Learning ◽

Wave Height ◽

Significant Wave Height ◽

Work Performance ◽

Feature Reduction ◽

Learning Approaches ◽

Significant Wave ◽

Wave Parameters ◽

Height Prediction

The prediction of wave height is one of the major problems of coastal engineering and coastal structures. In recent years, advances in the prediction of significant wave height have been considerably developed using flexible calculation techniques. In addition to the traditional prediction of significant wave height, soft computing has explored a new way of predicting significant wave heights. This research was conducted in the direction of forecasting a significant wave height using machine learning approaches. In this paper, a problem of significant wave height prediction problem has been tackled by using wave parameters such as wave spectral density. This prediction of significant wave height helps in wave energy converters as well as in ship navigation system. This research will optimize wave parameters for a fast and efficient wave height prediction. For this Pearson’s, Kendall’s and Spearman’s Correlation Coefficients and Particle Swarm Optimization feature reduction techniques are used. So reduced features are taken into consideration for prediction of wave height using neural network. In this work, performance evaluation metrics such as MSE and RMSE values are decreased and gives better performance of classification that is compared with existing research’s implemented methodology. From the experimental results, it is observed that proposed algorithm gives the better prediction as compared to PSO feature reduction technique. So, it is also concluded that Co-relation enhanced neural network is better as compared to PSO based neural network with increased number of features.

Download Full-text