Development of Time Series Model to Predict Daily Gold Price

Gold is ancient and one of the most precious and popular commodities in the world. Gold price forecasting is critical in financial decision-making, providing valuable information for investors in the gold market, sellers of gold items and stakeholders. Not much studies have been carried out in to forecast daily gold prices of Sri Lanka. The aim of this paper is to forecast the daily gold price rate (Rupees/troy ounce) using data from 2nd January 2018 to 14th June 2024 published by the Central Bank of Sri Lanka. The best fitted model was identified as ARIMA (1,1,1) + ARCH (2). The model was trained using data from 2nd January 2018 to 31st May 2024 and validated using data from the 3rd of June 2024 to 14th of June 2024. The model was statistically tested using standard statistical procedure and errors were found as white noise. The Mean Absolute Percentage Error (MAPE) for the training data set and validation data set were 0.748% and 1.002% respectively. The validation confirmed that the ARIMA (1,1,1) + ARCH (2) model effectively captures the dynamics of gold price movements, offering robust predictive power. These results indicate that the model is highly accurate and reliable for forecasting, making it a valuable tool for financial institutions and investors aiming to predict market trends and make informed investment decisions. Keywords: gold price rates; ARIMA models; Forecasting