Perancangan Aplikasi Penentuan Kualitas Bibit Arwana Menggunakan JST Backpropagation

Abstract

Super red Arowana is a popular ornamental fish. Arowana fish lovers will usually buy small arowana seeds (10-25 cm) to be raised and sold when they grow up. Many factors can be used to determine the quality of seeds such as the head, back, and tail of Arowana. But these factors are difficult to be see by some people. Therefore, to help the arowana lover who are still laymen, the authors make applications that can determine the quality of arowana seed based on the picture using artificial neural networks with backpropagation method. This research using case study design. Methods of design using prototyping model. The results showed that the value of the Mean Square Error (MSE) with 40 images obtained at 0.103 with the combination of parameters number of epochs 50, goal error 1e-1 / 1e-5, learning rate 0.01, number of neurons 50 and linear transfer function. Test results on 40 images with 28 training images and 12 test images with 2 output targets (good or poor seed quality) successfully identified 32 of the 40 images with a success rate of 80%. It can be concluded that the application produced reliable enough to determine the quality of arowana seeds.