Gold Price Prediction Based On Arima

Abstract

This study compares the forecasting performance of the Autoregressive Integrated Moving Average (ARIMA) model and the Long Short-Term Memory (LSTM) neural network in predicting daily closing prices of gold. Using a dataset covering the period from January 1, 2023, to May 25, 2025, both models were implemented in Python and evaluated using standard error metrics: Mean Absolute Error (MAE), Mean Squared Error (MSE), and Root Mean Squared Error (RMSE). The ARIMA model was constructed following differencing and optimal order selection based on the Akaike Information Criterion. The LSTM model employed a multi-layer architecture refined through experimental trials. Results indicate that the ARIMA model outperformed the LSTM network across all metrics, suggesting its greater suitability for modeling short-term gold price movements in this context.