COMPARISON OF THE ACCURACY OF ARIMA AND PROPHET MODELS IN PREDICTING THE STOCK PRICE OF PT INDOFOOD SUKSES MAKMUR

Stock price predictions are an important aspect in supporting investment decision-making, especially for large companies whose stock prices fluctuate dynamically. This study aims to compare the accuracy levels of the AutoRegressive Integrated Moving Average (ARIMA) and Prophet models in predicting the closing price of PT Indofood Sukses Makmur Tbk (INDF.JK) shares. The historical data used covers the period from January 1, 2019, to April 28, 2025, and was obtained from web scraping yahoo.finance.com. The research stages include data preprocessing, stationarity testing, training and testing data division, ARIMA and Prophet modeling, and performance evaluation using the Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE) metrics. The results show that the Prophet model with the parameter changepoint_prior_scale = 0.01 provides higher accuracy with an RMSE of 499.76, an MAE of 424.41, and an MAPE of 6.46% compared to ARIMA(1,1,0), which produces an RMSE of 907.50, an MAE of 701.36, and an MAPE of 9.98%. This study shows that Prophet produces better prediction performance for INDF.JK stocks, making it an alternative stock price forecasting model with an acceptable level of relative error.