IMPLEMENTASI GATED RECURRENT UNIT (GRU) UNTUK PREDIKSI HARGA SAHAM BANK KONVENSIONAL DI INDONESIA

Abstract

Advances in technology, information and technology at this time are growing rapidly, a lot of human work is facilitated by technology. Technology has also developed in investment instruments, especially in stock investments. Previously, when there was no technology, it would be very difficult to predict the stock price of state-owned banks in the future for ordinary people who do not understand fundamental or technical analysis. However, with technology, especially in the field of Deep Learning, it will be very possible to predict future stock prices without having to understand fundamental or technical analysis. In this study, stock price predictions of state-owned banks for the next 30 days were made using the Gated Recurrent Unit (GRU) model on stocks of state-owned banks in Indonesia, namely BRI, BNI, BTPN, and Mandiri bank shares. The data used uses historical close price data on the stock for 5 years with a total of 1260 data. From the results of the research for test data, the smallest RMSE value was found in BTPN bank shares of 23,91164 followed by BRI bank shares of 264,1475, BNI bank of 427,7984 and the largest RMSE value in BMRI bank shares amounting to 907,4804