Autoregressive Integrated Moving Average Model for the Forecasting of Gold Prices

Edios Merah (a*), Rini Novrianti Sutardjo Tui (a), Aryanti Virtanti Anas (a)

Autoregressive Integrated Moving Average Model for the Forecasting of Gold Prices
Edios Merah (a*), Rini Novrianti Sutardjo Tui (a), Aryanti Virtanti Anas (a)

(a) Mining Engineering Department, Faculty of Engineering, Hasanuddin University.
Jalan Poros Malino Km.6 Bontomarannu, Gowa, Sulawesi Selatan 92171, Indonesia.

Abstract

Gold prices are subject to significant fluctuations influenced by various economic, geopolitical, and market dynamics. Understanding these fluctuations is crucial for investors to make informed decisions. This research employs the Autoregressive Integrated Moving Average (ARIMA) model to forecast gold prices, considering its ability to analyze time series data effectively. The study utilizes historical gold price data from January 2008 to December 2022), to construct and validate the ARIMA model. The research methodology involves several stages, including data preprocessing, identification of ARIMA parameters, parameter estimation, diagnostic testing, and forecasting. Stationarity testing using the Augmented Dickey-Fuller (ADF) test reveals the non-stationarity of the data initially, which is addressed through differencing. Correlogram analysis aids in identifying appropriate ARIMA parameters, and diagnostic tests ensure the adequacy of the chosen model. After comparing and evaluating multiple ARIMA models, the ARIMA (22,1,25) model is selected as the most suitable for forecasting gold prices. This decision is based on criteria such as the Akaike Information Criterion (AIC), Hannan-Quinn Criterion, and R-squared values. The chosen model undergoes white noise testing to validate its suitability for forecasting.Using the validated ARIMA(22,1,25) model, gold price predictions are made for the period from January 2023 to December 2027. The predicted prices provide valuable insights for stakeholders, assisting in investment decisions, financial planning, and risk management strategies. Overall, this study demonstrates the efficacy of the ARIMA model in forecasting gold prices and underscores its importance in navigating the complexities of the gold market. By leveraging advanced statistical techniques and historical data analysis, stakeholders can gain a better understanding of future price trends and make more informed decisions in the gold market.

Keywords: Autoregressive Integrated Moving Average Model- Forecasting of Gold Prices- Identification of ARIMA Parameters.

Topic: Mineral and Energy Regulation and Economics