Prediksi Pasang Surut Air Laut Menggunakan Jaringan Syaraf Tiruan Backpropagation

Abstract

The tide of sea water has an effect on the activities carried out in the sea, namely shipping activities, fishing activities, and loading and unloading of ships, because tidal events occur not at the same time therefore the need for tide level prediction. Tide level data for forecast based historical tide data obtained from BMKG Tanjungpinang from January 1 to February 11, 2015, research was done by using Backpropagation. This study using as many as 1000 high tide data with some input parameters such as max iteration, target error, learning rate, number of input, and update learning rate. The accuracy of this forecast is measured by calculating the average error using MSE (Means Square Error). The best modeling result of Backpropagation with 5 hidden layer and learning rate 0,9 produce the smallest MSE 0,0035861.