New Adaptive Financial Transformer Enhances Stock Prediction Accuracy

Researchers have developed an Adaptive Financial Transformer (AFT) model designed to improve stock return prediction in volatile financial markets. This novel architecture incorporates a Market Regime Encoder and an Adaptive Gate Network to dynamically adjust self-attention mechanisms based on semantic relationships among financial indicators. The AFT model groups 95 engineered financial features into 11 categories and adapts attention according to latent market regimes, addressing issues like sequence alignment and backtesting inflation found in previous studies. Experiments show competitive predictive performance with reduced model complexity and improved parameter efficiency, offering an interpretable Transformer for financial time-series forecasting. AI

IMPACT Introduces a more interpretable and efficient Transformer architecture for financial time-series forecasting, potentially improving algorithmic trading strategies.

RANK_REASON The cluster contains an academic paper detailing a new model architecture and experimental results. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• Adaptive Financial Transformer
• Machine Learning
• Transformer

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →