PERBANDINGAN METODE ELMAN RECURRENT DAN RADIAL BASIS FUNCTION UNTUK PERAMALAN WISATAWAN

Istiqomah Ambarwati; Scolastika Mariani

doi:10.15294/ujm.v11i2.61066

Aan, A., Permana, J., Prijodiprodjo, W., Informatika, J. M., & Neural, E. R. 2014. Sistem evaluasi kelayakan mahasiswa magang menggunakan elman recurrent neural network. IJCCS (Indonesian Journal of Computing and Cybernetics Sistems). 8(1): 37–48.
Afrianty, I., Humairah, E., Sanjaya, S., Gusti, S. K., & Rouza, E. 2018. Penerapan jaringan syaraf tiruan elman recurrent neural network ( ernn ) untuk prediksi penjualan pilus. Prosiding Seminar Nasional Teknologi Informasi, Komunikasi Dan Industri. Pekanbaru 158–164 November 2018.
Aman, Z., Ezzine, L., Fakhradine, Y., Bahi, E., & Moussami, H. E. L. 2019. Improving the modeling and forecasting of fuel selling price using the radial basis function technique : A case study. J. Algorithms & Computational Technology. 13: 1–9.
Aprianti, W., Permadi, J., & Rhomadhona, H. 2020. Perbandingan elman recurrent neural networks, backpropagation neural networks, dan exponential smoothing dalam peramalan produksi palawija. J. Mathematics Education, Science and Technology. 5(2): 206–220.
Caselli, M., Trizio, L., Gennaro, G., & Ielpo, P. 2008. A simple feedforward neural network for the PM10 forecasting : comparison with a radial basis function network and a multivariate linear regression model. Water,Air, Soil ,Pollut. 201(1): 365–377.
Chang, F., Liang, J., & Chen, Y. 2001. Flood forecasting using radial basis function neural networks. Transactions On Systems, Man, And Cybernetics—Part C: Applications And Reviews. 31(4): 530–535.
Cynthia, E. P., Yanti, N., Fitriani, Y., & Yusuf, M. 2016. Penerapan metode Elman Recurrent Neural Network ( ERNN ) untuk peramalan penjualan. J. Education Informatic Technology and Science. 1(2): 49–61.
Dinzi, R., & Energy, A. W. 2020. The use of meteorology data in short-term prediction of wind speed for wind turbine using elman recurrent neural network. International Conference on Data Science, Artificial Intelligence, and Business Analytics (DATABIA). Medan 93–98 Oktober 2020.
Fauzannissa, R. A., Yasin, H., & Ispriyanti, D. 2015. Peramalan harga minyak mentah dunia menggunakan metode radial basis function neural network. J. Gaussian. 5(1): 193–202.
Harsono, I. T., Wibowo, A. T., & Dayawati, R. N. 2011. Analisis dan implementasi elman recurrent neural network dan tabu search pada prediksi harga perak. Tugas Akhir. Bandung: Fakultas Teknik Informatika Telkom.
Lin, G., & Chen, L. 2005. Time series forecasting by combining the radial basis function network and the self-organizing map. Hydrological Process. 19: 1925–1937.
Talahatu, J., Benarkah, N., & Jimmy. 2015. Penggunaan aplikasi sistem jaringan syaraf tiruan berulang elman untuk prediksi pergerakan harga saham. J. Ilmiah Mahasiswa Universitas Surabaya. 4(1): 1–12.
Tarigan, R., Yasin, H., & Prahutama, A. 2018. Peramalan Indeks Harga Saham Gabungan (IHSG) dengan metode radial basis function neural network menggunakan GUI Matlab. J. Gaussian. 7(4): 431–442.
Wang, J., & Wang, J. 2016. Forecasting energy market indices with recurrent neural networks : Case study of crude oil price fluctuations. J. Energy. 102(2016): 365–374.
Wiyanti, D. T., & Pulungan, R. 2013. RBF and ARIMA combined for time series forecasting. Seminar Nasional Aplikasi Teknologi Informasi (SNATI). Yogyakarta 22-27 Juni 2013.

Abstract viewed - 145 times
PDF downloaded - 55 times

Affiliations

Istiqomah Ambarwati
Universitas Negeri Semarang

Scolastika Mariani
Universitas Negeri Semarang

Content Top

PERBANDINGAN METODE ELMAN RECURRENT DAN RADIAL BASIS FUNCTION UNTUK PERAMALAN WISATAWAN

Istiqomah Ambarwati ,
Scolastika Mariani

Vol 11 No 2 (2022)

DOI 10.15294/ujm.v11i2.61066

Submitted: Oct 9, 2022

Published: Jul 7, 2023

Abstract

Elman Recurent Neural Network (ERNN) adalah model jaringan syaraf yang memiliki minimal satu feedback loop, sedangkan Radial Basis Function Neural Network (RBFNN) adalah model JST yang mentransformasikan input secara nonlinier menggunakan fungsi aktivasi Gaussian pada lapisan tersembunyi sebelum diproses secara linier di lapisan output. Tujuan penelitian untuk memperoleh model terbaik dari ERNN dan RBFNN dengan fungsi pelatihan terbaik untuk peramalan jumlah kunjungan wisatawan mancanegara di Indonesia. Hasil ERNN model terbaik diperoleh pada arsitektur jaringan 2-15-1 dengan algoritma pelatihan gradient descent with momentum and adaptive learning rate dengan momentum 0,2 , learning rate 0,1 , nilai MSE dan MAPE pengujian sebesar 1,70E+10 dan 13,5869%, serta akurasi jaringan sebesar 86,4131%. Sedangkan hasil peramalan menggunakan RBFNN diperoleh model terbaik pada arsitektur jaringan 2-8-1 dengan algoritma pelatihan gradient descent with momentum and adaptive learning rate atau gradient descent with momentum dengan spread 1. Nilai MSE dan MAPE pengujian sebesar 1,26E-02 dan 6,7043%, serta akurasi jaringan sebesar 93,2957%. Model terbaik untuk peramalan jumlah kunjungan wisman adalah RBFNN (2-8-1).

Elman Recurent Neural Network (ERNN) is a neural network model that has at least one feedback loop, while Radial Basis Function Neural Network (RBFNN) is an ANN model that transforms input nonlinearly using a Gaussian activation function in the hidden layer before being processed linearly in the output layer. The purpose of the study was to obtain the best model from ERNN and RBFNN with the best training fuction for forecasting the number of foreign tourist arrivals in Indonesia. The best ERNN model results are obtained on a 2-15-1 network architecture with a gradient descent with momentum and adaptive learning rate training algorithm with a momentum of 0.2, learning rate of 0.1, MSE and MAPE testing values 1.70E+10 and 13,5869%, and network accuracy of 86,4131%. While the results of forecasting using RBFNN obtained the best model on the network architecture 2-8-1 with a training algorithm gradient descent with momentum and adaptive learning rate or gradient descent with momentum with a spread of 1. The MSE and MAPE testing values 1.26E-02 and 6, 7043%, and network accuracy of 93.2957%. The best model for forecasting the number of foreign tourists visiting is RBFNN (2-8-1).