Prediksi Tren Pergerakan Harga Saham PT Bank Central Asia Tbk, Dengan Menggunakan Algoritma Long Shot Term Memory (LSTM)

Abstract

Shares are valuable documents that prove ownership of a company. Stock investment is one of the right choices to get more profit. There are various stocks in Indonesia, one of which is the shares of PT Bank Central Asia Tbk (BBCA). However, in making stock investments, it is necessary to analyze the data of a company that can determine the increase or decrease in a stock price. Very dynamic movements require data modeling to predict stock prices in order to get a high level of accuracy. In this study, modeling using the Long-Short Term Memory (LSTM) algorithm to predict BBCA stock prices. The data used is secondary daily data obtained from securities with a date range of January 3, 2011 to December 30, 2022. The main objective of this research is to analyze the accuracy of the LSTM algorithm in forecasting stock prices and to analyze the number of epochs in the formation of the optimal model. The optimal epoch variation is obtained with the number of epochs of 5 and batch size 1. The resulting values include Mean Absolute Error (MAE) of 96.92, Mean Squared Error (MSE) of 16185.22 and Root Mean Squared Error (RMSE) of 127.22. The results of this study provide further insight into the performance of the LSTM algorithm in stock price prediction and show that with the right parameter settings, it can be a useful tool for investors in making better investment decisions