基于机器学习的斜坡堤越浪量预测方法研究

针对斜坡堤越浪量预测方法,分别建立集成神经网络(ensemble neural network,ENN)、随机森林(random for-eset,RF)和支持向量回归机(suppport vector regression,SVR)3种机器学习模型对斜坡堤越浪量进行预测,并利用决定系数R2和均方根误差RMSE来评估模型性能.最后,对3种模型的性能进行分析.结果显示,集成神经网络模型的决定系数R2和均方根误差RM S E分别约为0.96和0.0018,随机森林模型的决定系数R2和均方根误差RMSE分别约为0.97和0.0014,支持向量回归机模型的决定系数R2和均方根误差RMSE分别约为0.94和0.002.对比发现,3种模型的决定系数都达到0.9以上,都具有较高的预测精度,随机森林相比其他两个模型精度更高.

Research on Prediction Methods of Sloping Breakwater Overtopping Based on Machine Learning

Aiming at the method of predicting the overtopping on sloping breakwater, this paper establishes three machine learning model pairs: ensemble neural network (ENN), random forest ( RF), and support vector regression (SVR). Predicts the amount of wave over the sloping breakwater. And uses the coefficient of determination (R2) and the root mean square error (RMSE) to evaluate the performance of the model. Finally, the performance of the three models is analyzed. The results show that the coefficient of determination (R2) and the root mean square error (RMSE) of the in ensemble neural network model are about 0.96 and 0.0018, respectively. The coefficient of determination (R2) and root mean square error (RMSE)of the random forest model are about 0.97 and 0.0014, respectively. The coefficient of determination (R2)and the root mean square error (RMSE)are about 0.94 and 0.002, respectively. It is found by comparison that the determination coefficients of the three models are all above 0.9, and they all have higher prediction accuracy. Random forest is better than the other two.