Exploring IDX Composite Growth Using LSTM Model With US Economic Indicators

The purpose of this study is to analyze the IDX Composite’s growth, trends, and fluctuations using a machine learning approach, especially the Long Short Term Memory (LSTM) model. Besides to that model, the GRU, VAR, and SARIMA models will be used to assess the accuracy of all models examined in this study. LSTM was chosen as the main model in this study because it can capture recurring and fluctuating stock movement patterns. However, aside from utilizing the LSTM analysis model, this study also considers US economic indicators, including the interest rate, unemployment rate, and inflation rate, as external factors influencing the IDX Composite movement. The data used includes historical series of the IDX Composite from 1990 to 2024 as well as data on US economic indicators throughout the same time period in order to provide the analysis an established basis. The results of this study demonstrate that LSTM performs better than other models in terms of accuracy and effectively captures changes in stock prices and market trends in the IDX Composite. Thus, the use of LSTM in this research proves that LSTM demonstrates its potential to be a reliable analytical tool in modeling stock price movements in the IDX Composite while taking into consideration the impact of global economic changes.