Anfis Model for Prediction of Current Velocity at Filyos Region

Authors : Rıfat Tür, Dilayda Soylu

Pages : 1-9

Doi:10.24107/ijeas.281420

View : 19 | Download : 12

Publication Date : 2016-12-26

Article Type : Research Paper

Abstract :Current velocity plays a significant role in coastal engineering, especially coastal sedimentation, coastal pollution transmission, and design of coastal structures. Moreover, it is great important to determine coastal pollution propagation in time and the area affected by pollution transmission. Because of these reasons, current velocity is predicted based on observed data in this study. Current velocity data which are measured for 2 hours during 2 years in Filyos Region are utilized to develop several Adaptive Neuro-Fuzzy Inference System insert ignore into journalissuearticles values(ANFIS); models on Matlab to estimate future current velocity. After prediction of two hourly averages of current velocities from previous values by ANFIS model, the predicted data is compared with the actual one measured in the field. Therefore, statistical parameters in literature including root mean square error insert ignore into journalissuearticles values(RMSE);, mean absolute error insert ignore into journalissuearticles values(MAE);, and correlation coefficient insert ignore into journalissuearticles values(R); are used to test acceptability of proposed ANFIS models. The study results indicate that proposed models provide better results in comparison to widespread stochastic approaches. Consequently, this study is an alternative to other prediction methods considering the aims of current velocity prediction mentioned above.
Keywords : Current velocity forecasting, Adaptive Neuro Fuzzy Inference System ANFIS, Filyos Region