IMPLEMENTASI JARINGAN SYARAF TIRUAN MENGGUNAKAN METODE PROPAGASI BALIK UNTUK MEMPREDIKSI TINGKAT KEMACETAN DI KOTA PALEMBANG
Daftar Isi:
- During this time in heavy traffic, so many different forms of research in predicting with the various types of parameters used. The purpose why take this congestion problem to help in predicting the congestion that often occurs in big cities by using the minimum error variations that is in the Backpropagation method. Backpropagation artificial neural networks are one form of supervised learning, using the standard Backpropagation training process. In the Backpropagation method it has many layers (multilayer network), among others, the input layer (input layer) handles 1 piece, consisting of 1 to n input units, hidden layer (minimum layer 1), consisting of 1 to p hidden units and layers resulting (output layer) have 1 piece, consisting of 1 to m output units. In Backpropagation, there is also an intensive application process for weight renewal that will be used at every age. In this study, testing will be carried out in 2 ways, testing with Backpropagation standards and testing with minimum error variations. In testing with the Backpropagation standard taken as 36 sample data for testing with a minimum error parameter 3 and a maximum of epoch 5 resulting in an accuracy rate of 97% with predictive data that matches the target data increase in 35 precise data. While with variations using the minimum error, when the minimum error of 7 produced a fairly low accuracy of 43.7% and increased significant when the minimum error at 5 and 3 with the resulting accuracy of 89.6% and 87.8%. With the change of minimum error, the result accuracy also varies. Then the change of minimum error affected to the level of accuracy accuracy. By using one of the minimum variations of this error, it can provide fairly variable level results, so that it can be seen the level of accuracy in predicting the level of congestion.