Prediksi IHSG dengan Backpropagation Neural Network

Andy Santoso (Universitas Multimedia Nusantara)
Seng Hansun (Universitas Multimedia Nusantara)

Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi)
Vol 3 No 2 (2019): Agustus 2019

IDX Composite is a combination of all common stock and preferred stock which registered on Bursa Efek Indonesia (BEI). IDX Composite is often used by investor to predict the stock price to get profit. But, to predict the stock price is not easy, hence it yields a high risk to investor. This study offers the usage of backpropagation algorithm to minimize the risk. Backpropagation is a supervised algorithm and will be made in Python programming language, in this case, backpropagation will use and learn the past 5 days data to predict the outcome. Also, this study shows that backpropagation have a high accuracy which reflects in Mean Square Error Testing value of 320.49865083640924 to predict IDX Composite using 0.3 learning rate and 3000 epoch.