Article Per Year (5 Year)
p-Index From 2021 - 2026
0
P-Index
This Author published in this journals
All Journal Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer
Grady Davinsyah
Fakultas Ilmu Komputer, Universitas Brawijaya
Computer Science & IT Control & Systems Engineering Education Electrical & Electronics Engineering Engineering

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Pelatihan Multi-Layer Neural Network Menggunakan Algoritma Genetika untuk Memprediksi Harga Saham Esok Hari (T+1) Grady Davinsyah; Wayan Firdaus Mahmudy; Rizal Setya Perdana
Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer Vol 2 No 6 (2018): Juni 2018
Publisher : Fakultas Ilmu Komputer (FILKOM), Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (1017.541 KB)

Abstract

Stock is one of investment instrument which is popular because of the high profit potential and risk. These profit potential and risk are caused by fluctuations of the stock price in the stock market. To minimalize the risk, a system which is able to predict closing price of the next day is required. The architecture which is used in this research is multi-layer neural network. This architecture is trained with 2 different training methods, which is backpropagation and genetic algorithm. Both of the methods aim to gain weights of all network's architecture. Backpropagation's parameters which obtained during the research are 4500 iteration and 0.7 learning rate. For genetic algorithm's parameters which obtained during the research are 2000 generations, population size of 200, crossover rate 0.1 and mutation rate 0.9. By using those parameters, average RMSE value which produced using backpropagation algorithm is 0.048006. Meanwhile when using genetic algorithm as a training method, average RMSE value which produced by the network is 0.065205. So in this research, average error value which is produced by using backpropagation training is smaller than using genetic algorithm training method.
Co-Authors Rizal Setya Perdana Wayan Firdaus Mahmudy