Hatuwe, Razif Zulvikar and Findawati, Yulian (2026) The Use of Facebook Prophet in Predicting Stock Price. International Journal on Human-Computing Studies, 8 (1). ISSN 2615-8159

Text IJHCS_5769+_Razif+Zulvikar+Hatuwe_The+Use+of+Facebook+Prophet+in+Predicting.pdf Download (501kB)

Abstract

This study examines the effectiveness of the Facebook Prophet model in predicting stock prices. The data consist of daily stock prices obtained from Yahoo Finance from July 30, 2019 to July 30, 2024, totaling 1,214 observations. A quantitative time series approach was applied by testing several Prophet configurations, including the basic model, the addition of external regressors, parameter tuning, and their combination. Model performance was evaluated using Mean Absolute Error (MAE) and Root Mean Square Error (RMSE). The best results were achieved by the model combining regressors with optimized parameters, producing an MAE of 36.31 and an RMSE of 52.06. Overall, the model provided reasonably accurate predictions for both short- and long-term periods, although bias increased with longer forecast horizons. These findings suggest that Facebook Prophet can be used as a supporting tool for stock price trend analysis based on historical data.

Item Type:	Article
Uncontrolled Keywords:	Facebook Prophet, stock-price forecasting, Time series, MAE, RMSE
Subjects:	H Social Sciences
Depositing User:	admin eprints
Date Deposited:	19 Jul 2026 11:45
Last Modified:	19 Jul 2026 11:45
URI:	http://eprints.umsida.ac.id/id/eprint/16811

Actions (login required)

View Item