From Index to Equity: Pre-Training Transformers for Stock Return Prediction
Researchers have developed a transformer-based model for stock return prediction, utilizing pre-training on a market index to enhance performance. The model, pre-trained on the Toronto Stock Exchange Index (TSX) and then fine-tuned on individual stocks, showed improved loss metrics compared to baseline models. While achieving lower mean squared error in regression tasks, its ability to generate higher average daily returns was surpassed by ensemble and XGBoost models. A practical application was also created to provide real-time trading support. AI
IMPACT Demonstrates potential for advanced ML techniques to refine quantitative finance strategies and trading tools.