FORECASTING SAHAM SYARIAH DENGAN MENGGUNAKAN LSTM

Abstract

Islamic stocks as one of the many stocks listed on the JCI are a barometer of the Islamic market in Indonesia. One approach in predicting stock prices is by using machine learning. The purpose of this study is to make a model that is used to predict JII shares using the LSTM approach. The data used amounted to 1402 records related to the Jakarta Islamic Index (JII) stock from March 4, 2014 - January 2, 2019. Model making uses 3 Epochs (1, 10 and 20). The results showed the best model was to use 20 Epochs. The increase in Epoch affects the value of MSE and RMSE which are getting smaller. For Epoch 20, the values of MSE and RMSE are ~ 0.00019 and ~ 0.014, respectively.