PERBANDINGAN RADIAL BASIS FUNCTION DAN RECURRENT NEURAL NETWORK PADA PREDIKSI CURAH HUJAN DI PALEMBANG

Main Authors: SIAGIAN, DEWI PUTRI, Sazaki, Yoppy, Saputra, Danny Matthew
Format: Thesis NonPeerReviewed Book
Bahasa: ind
Terbitan: , 2019
Subjects:
etc
Online Access: http://repository.unsri.ac.id/23251/1/RAMA_55201_09021281320004.pdf
http://repository.unsri.ac.id/23251/2/RAMA_55201_09021281320004_TURNITIN.pdf
http://repository.unsri.ac.id/23251/3/RAMA_55201_09021281320004_0006067406_0010058507_01_front_ref.pdf
http://repository.unsri.ac.id/23251/4/RAMA_55201_09021281320004_0006067406_0010058507_02.pdf
http://repository.unsri.ac.id/23251/5/RAMA_55201_09021281320004_0006067406_0010058507_03.pdf
http://repository.unsri.ac.id/23251/7/RAMA_55201_09021281320004_0006067406_0010058507_04.pdf
http://repository.unsri.ac.id/23251/8/RAMA_55201_09021281320004_0006067406_0010058507_05.pdf
http://repository.unsri.ac.id/23251/9/RAMA_55201_09021281320004_0006067406_0010058507_06.pdf
http://repository.unsri.ac.id/23251/10/RAMA_55201_09021281320004_0006067406_0010058507_07_ref.pdf
http://repository.unsri.ac.id/23251/11/RAMA_55201_09021281320004_0006067406_0010058507_08_lamp.pdf
http://repository.unsri.ac.id/23251/
Daftar Isi:
  • Abstract-Rainfall plays a major role in economic growth of every region. Its predetermined and accurate are very important to many sectors especially to agriculture in Palembang. In this paper, two types of artificial neural networks (ANNs), Radial Basis Function (RBF) and Recurrent Neural Network (RNN) were used to predict the monthly rainfall values based on the data collected from Agency for Meteorology, Climatology and Geophysics in Kenten district (Palembang). Moreover, accuracy and root mean square error (RMSE) are the performance indices used for the comparative analysis. Experimental results showed that the recurrent neural network acts better prediction model with higher accuracy values but have 0,07 higher root mean square error values than radial basis function. Furthermore, we should use this method to predict the future data based on previously collected data.