ONE-WEEK WAVE PREDICTION USING GWM AND XGBOOST

Abstract

 The recent studies introduce machine learning-based wave prediction models using the Group Method of Data Handling (GMDH) and Artificial Neural Network (ANN) for one-week wave prediction along nearshore coasts in Japan because global wave forecast models predict waves on large spatial resolutions, being unreliable for nearshore wave predictions. The current study develops the GWM to XGBoost nearshore wave prediction model that transforms global wave model GWM forecast waves to nearshore ones. The XGBoost method is one of the machine learning techniques that the ensemble training method improves the discrimination efficiency by combining multiple-decision trees. Then, the study discused the accuracy of the GWM to XGBoost nearshore wave prediction model, and showed a good performance.