Monthly Gold Price Forecasting Using ANN And ARIMA

Abstract

Forecasting is a prominent statistical area with several applications, notably in econometrics. Many governments use it to set long-term objectives and make future choices. Establishing a reliable gold price model is critical because gold prices have a significant effect on the financial decisions of people, organizations, and different countries. This research examines the two primary forecasting methodologies in order to determine the optimum forecasting model for monthly gold prices in US dollars per one ounce of gold. The first strategy, known as Box-Jenkins, use the Autoregressive Integrated Moving Average (ARIMA) model, while the second employs the Feed Forward Neural Network (FFNN) model. The information is derived from the official website of IndexMundi, and it covers monthly gold prices in US dollars per one ounce of gold were used from January 2010 to December 2022. Alyuda NeuroIntelligence, R, and SPSS were utilized for analysis. This comparison also includes Akaike Information Criteria (AIC), Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and R2. The results show that the FFNN model fits better than the ARIMA model. Furthermore, because of lower MAE, RMSE, and AIC values, the FFNN model has fewer errors than the ARIMA model and is much better in terms of goodness of fit.