Leveraging Artificial Intelligence for Gold Price Forecasting: An LSTM-Based Analysis

The predictive power of Artificial Intelligence (AI) is revolutionising financial market analysis. This paper leverages a sophisticated branch of AI deep learning to forecast the price of gold, a critical safe-haven asset. We implement a Long Short-Term Memory (LSTM) neural network, an architecture renowned for processing sequential data, to model daily gold prices from January 1993 to April 2024. After a comprehensive exploratory analysis, the data were normalised and transformed into time-series samples. The constructed LSTM model was trained to identify complex, non-linear patterns in the historical data. The model demonstrated exceptional performance, achieving a remarkably low mean squared error, confirming its capability to accurately learn temporal dependencies. The finalised AI model was then deployed to generate a forecast for gold prices over a 60-day horizon. The results strongly indicate that AI-driven models, particularly LSTM networks, offer a powerful and reliable framework for forecasting volatile financial instruments like gold, providing valuable insights for investors and policymakers. Importantly, this study focuses solely on an LSTM-based deep learning approach rather than broader ensembles of machine learning methods.