An Effective Fuzzy Feature Selection and Prediction Method for Modeling Tidal Current: A Case of Persian Gulf

Abstract

This paper develops a new two-stage approach for accurate modeling and prediction of tidal current. The proposed method makes use of a novel fuzzy feature selection to extract the most preferable features from the tidal current speed and direction data set. The selected features are further used to train a support vector regression for accurate prediction. The setting parameters of the proposed model are trained by a new optimization algorithm based on the harmony search algorithm to get to the most optimal training targets. The proposed optimization algorithm makes use of the crossover and mutation operators from genetic algorithm to escape from the local optima and find the global solutions. Experimental tidal data from Persian Gulf, Iran, are used to assess the accuracy and performance of the proposed model. The results show the appropriate performance and high precision of the proposed model in comparison with other famous methods.