Comparing the Best-Fit and Integrating Generative Deep Learning with Time Series Models for Metal Price Forecasting: A Predictive Analysis

Forecasting metal prices in volatile financial markets is a challenging task due to the non-linear and stochastic nature of price movements. This research integrates deep learning techniques with time series models to predict metal prices, focusing on the MCX Index. A novel Generative Adversarial Network (GAN) framework, coupled with a Convolutional Neural Network (CNN) as a discriminator and Gated Recurrent Units (GRU) as a generator, is proposed. Historical NSE metal spot price data, specifically gold and silver, were used for model training and evaluation. The empirical results demonstrated that the proposed GAN-based architecture outperformed other deep learning models, including LSTM and Bi-LSTM, in terms of forecasting accuracy. The study provides valuable insights into the application of deep generative models for financial forecasting, addressing key challenges associated with modeling non-stationary and noisy financial data.