ARIMA Model Selection for Composite Stock Price Index in Indonesia Stock Exchange

Abstract

Composite Stock Price Index (CSPI) can be used as a reflection of the national economic condition of a country because it is an indicator to know the development the capital market in a country. Therefore, the movement in the future needs to be forecast. This study aims to build a model for the time series forecasting of Indonesia Composite Index (ICI) using the ARIMA model. The data used is the monthly data of ICI in Indonesia Stock Exchange (IDX) from January 2000 until December 2017 as many as 216 data. The method used in this research is the Box-Jenkins method. The autocorrelation (ACF) and partial autocorrelation function (PACF) are used for stationary test and model identification. The maximum estimated likelihood is used to estimate the parameter model. In addition, to select a model then used Akaike’s Information Criterion (AIC). Ljung-Box Q statistics are used for diagnostic tests. In addition, to show the accuracy of the model, we use Root Mean Squared Error (RMSE), Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE) and the most appropriate model is ARIMA (0, 1, 1).

Study of India stock prices (2023) of Banks and prediction based on ARIMA models

“…Suresh Kumar et al [11] used the Nonlinear Autoregressive Distributed Lag (NARDL) model using weekly data over the period from January 1997 to June 2019 to examine the relationship between the price of natural gas, crude oil, gold, exchange rate, and stock market index in the Indian setting. Numerous investigations are connected to the ARIMA time series model, see in [12][13][14][15][16][17].…”