Optimal ANFIS Model for Forecasting System Using Different FIS

Abstract

Adaptive Network Based Fuzzy Inference System (ANFIS) with time series analysis is one of the intelligent systems that can be used to predict with good accuracy in all fields like in meteorology's field. However, some researches on prediction or forecasting do not emphasize the structure of the ANFIS fuzzy inference system. Thus, in this research, the optimization of the ANFIS model for forecasting maritime weather was carried out by analyzing the exact initialization determination in the three ANFIS fuzzy inference structures, they are grid partition, subtractive clustering, and fuzzy cmeans clustering. The input variable used in this research is each variable for the previous two hours, one hour, and at a time and the output of this system is the prediction of one hour, six hours, twelve hours and one day ahead of the ocean current speed (cm/s) and wave height (m) using all three FIS ANFIS approaches. Based on the results, the smallest goal error (RMSE and MSE) of the three FIS ANFIS approaches used for predicting the current speed and wave height, the best model is produced by Subtractive Clustering. It can be seen that Subtractive Clustering produces the smallest RMSE and MSE error values.