<![CDATA[Stock price movements are highly volatile, requiring reliable forecasting models to support investment decision-making. This study compares the performance of time-series models, namely the Autoregressive Integrated Moving Average (ARIMA) and the Prophet, in predicting the closing price of PT Astra International Tbk (ASII.JK). The study employs secondary data obtained from Yahoo Finance covering the period from October 2020 to October 2025. Model evaluation is conducted using an out-of-sample backtesting scheme with RMSE, MAE, MAPE, and directional accuracy (DA) as performance metrics. The results indicate that the ARIMA(2,1,2) model provides superior numerical accuracy, achieving a MAPE of 6.26%, while the Prophet model with a changepoint prior scale of 0.5 yields a MAPE of 7.33%. In terms of price movement direction, Prophet demonstrates a higher DA value of 57.26%. Visual analysis shows that ARIMA predictions closely track actual price movements, with relatively small deviations, whereas Prophet produces increasingly wide uncertainty intervals at longer forecasting horizons. Based on these findings, ARIMA is more suitable for precise price forecasting, while Prophet is better suited for analyzing price direction and trend dynamics.]]>
Copyrights © 2026
