Forecasts marine weather on Java sea using hybrid methods: TS-ANFIS

Indonesia is an archipelago. Consequently, the majorities are working around the sea such as a fisherman. While the number of activities at sea are increasing more accident occurred are rising. This research presents marine weather prediction system using Hybrid Methods TS-ANFIS (Adaptive Neuro Fuzzy Inference System — Time Series) in order to anticipate bad weather and reduce risk. This method use both ocean current and wave height at Java Sea particularly on Gresik in order to forecast ocean current velocity and wave height. Input variables used in this paper are data at (t), an hour before (t-1), and two hours before (t-2) and obtained next hour, next 6 hours, next 12 hours, and next day prediction as output. The results indicate that ocean current speed attain 16.97327 cm/s; 13.22302 cm/s; 10.21107 cm/s; 14.09871 cm/s with mean error is about 0.12993; 1.5758; 1.3182; 0.82613 while wave height reach 0.45554 m; 0.48286 m; 0.46395 m; 0.54571 m with mean error is about 0.0012247; 0.018619; 0.046584; 0.060206. Therefore, it was safe to sailing on 1st January 2016.