This project implements a complete stock price prediction pipeline using multiple AI techniques. It employs a Generative Adversarial Network (GAN) with LSTM as the generator and CNN as the discriminator for time-series forecasting. Hyperparameter optimization is handled via Bayesian optimization with Gaussian processes and reinforcement learning (Rainbow, PPO). Additional inputs include BERT-based sentiment analysis, Fourier transforms, stacked autoencoders, and ARIMA for comprehensive market analysis.