Automated detection of offshore wave power using machine learning techniques

•More than 206.000 wave power data are processed to achieve high accuracy using the advantage of deep learning-based models.•A deep learning based RBM, ELM and Regression models are used to predict wave power.•Feature selection is performed with Relief algorithm to improve prediction accuracy.•Input weights and input bias analyzed in RBM are used as parameters in calculation of ELM again.•The best accuracy value for RBM and RBM-ELM methods are obtained with R-Square values as 81.76% and 98.81%, respectively.