Workflow
AI概念指数轮动模型
icon
Search documents
文本选股策略超额收益收窄
HTSC· 2025-12-28 11:32
Group 1 - The LLM-FADT text stock selection strategy has underperformed relative to the CSI 500 index by -1.5% this month, with a year-to-date excess return of 2.9% [1][18] - The LLM-FADT strategy has an annualized return of 29.05% since January 2017, with an excess annualized return of 26.56% relative to the CSI 500 index, and a Sharpe ratio of 1.13 [1][20] - The AI industry rotation model recommends holding non-bank financials, petrochemicals, beverages, steel, and utilities for the upcoming week, with an annualized return of 26.49% since 2017 [3][41] Group 2 - The AI theme index rotation model suggests holding petrochemical industry, 300 non-bank, Shenzhen dividend, and China construction indices for the next week, with an annualized return of 16.58% since 2018 [4][28] - The AI concept index rotation model has an annualized return of 22.29% since 2018, with a maximum drawdown of 19.19% [33][34] - The AI industry rotation model has a year-to-date excess return of 29.13% and an annualized excess return of 19.53% [39][41] Group 3 - The AI-enhanced stock selection strategy based on full-frequency fusion factors has achieved a year-to-date excess return of 19.98% relative to the all-A equal-weight benchmark [2][6] - The AI-enhanced CSI 1000 strategy has a year-to-date excess return of 24.24%, with an annualized excess return of 21.89% since 2017 [2][10] - The LLM-FADT strategy has shown more stability and smaller excess drawdowns compared to the BERT-FADT strategy since October 2024 [18][21]