PEMANFAATAN JARINGAN SARAF TIRUAN PROPAGASI BALIK UNTUK MODEL PREDIKSI DERET WAKTU PASANG SURUT

Afif Widaryanto (Badan Pengkajian dan Penerapan Teknologi)
Fineza Ilova (Badan Pengkajian dan Penerapan Teknologi)

Jurnal Riset dan Rekayasa Kelautan
Vol. 1 No. 2 (2020): Oseanika: Jurnal Riset dan Rekayasa Kelautan - Desember 2020

[The Use of Backpropagation Neural Network for Time Series Tide-Level Prediction Model] Most of ocean related activities require tidal prediction data. This requires a prediction system with high accuracy. The widespread application of artificial intelligence with its various reliability inspired this research to apply tidal prediction models using artificial neural networks. With the input of tidal data for the previous seven days to predict the tide for the next 6 and 12 hours, it can be modeled using an artificial network based on back propagation learning method. As a result, the performance of the prediction model testing is very satisfying with an average accuracy of above 90% and low MSSE (mean sum square error) values.Keywords: tide; artificial intelligent; neural network; prediction model