Prediksi Hasil Panen Udang Vaname menggunakan Algoritme Backpropagation Neural Network

Harris Imam Fathoni (Fakultas Ilmu Komputer, Universitas Brawijaya)
Bayu Rahayudi (Fakultas Ilmu Komputer, Universitas Brawijaya)
Dian Eka Ratnawati (Fakultas Ilmu Komputer, Universitas Brawijaya)

Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer
Vol 6 No 8 (2022): Agustus 2022

Shrimp (Caridea) is one of the commodities that are often exported every year. In cultivation, to maintain the productivity of shrimp harvests, there are several parameters that must be seen every day. Forecasting is used to help farmers to be used as a reference in terms of shrimp maintenance until the final harvest. This research used shrimp cultivation for four years from 2016 until 2019 as data inputs Then normalization of data is carried out which will then be processed using the backpropagation method to find the forecasting value for vaname shrimp yields. The final results were denormalized again and then evaluated using the MAPE (Mean Absolute Percentage Error) method and obtained in this study the optimal value with a minimum MAPE of 3.65% using the hidden neuron parameter = 1, the epoch value = 1000 and the initial random weight range value from -1 up to 1.