StockGPT: A GenAI Model for Stock Prediction and Trading
Dat Mai showcases a generative AI innovation in stock market prediction with ‘StockGPT: A GenAI Model for Stock Prediction and Trading’, an autoregressive model trained on extensive historical U.S. stock return data.
- Utilizes an attention mechanism that identifies hidden patterns in stock returns for prediction.
- Achieved remarkable results in a test sample with portfolios, boasting a 119% annual return and a Sharpe ratio of 6.5 from 2001-2023.
- StockGPT-based portfolios eclipse traditional momentum and reversal strategies, as well as encompassing major stock market factors.
- Represents a major leap forward for generative AI in financial decision-making, outstripping traditional, manual strategies in complexity and performance.
StockGPT’s success exemplifies the disruptive potential of AI in finance, offering an edge in market analysis and investment strategies. It denotes a shift towards data-driven, AI-augmented trading that could dominate future financial markets and warrants close monitoring and further exploration for potential refinement and broader application.
Personalized AI news from scientific papers.