量化选股策略周报：本周指增组合超额回撤-20251025

Core Insights - The report highlights the construction of an AI-based low-frequency index enhancement strategy using deep learning frameworks to build alpha and risk models [4][16] - The market indices showed positive performance as of October 24, 2025, with the Shanghai Composite Index rising by 2.88%, the Shenzhen Component Index by 4.73%, and the CSI 300 by 3.24%, indicating a market uptrend despite reduced trading volume [6][9][10] Market Index Performance - As of October 24, 2025, the Shanghai Composite Index reached 3950.3 points, with a weekly increase of 2.88% and a year-to-date increase of 17.86% [10] - The Shenzhen Component Index stood at 13289.2 points, increasing by 4.73% weekly and 27.60% year-to-date [10] - The CSI 300 Index was at 4660.7 points, with a weekly rise of 3.24% and a year-to-date increase of 18.44% [10] Index Enhancement Fund Performance - As of October 24, 2025, the CSI 300 index enhancement fund had a minimum excess return of -1.29%, a median of -0.08%, and a maximum of 1.86% for the week [13] - The CSI 500 index enhancement fund reported a minimum excess return of -1.78%, a median of 0.02%, and a maximum of 1.07% [13] - The CSI 1000 index enhancement fund had a minimum excess return of -1.39%, a median of 0.29%, and a maximum of 1.36% [13] Year-to-Date Performance of Index Enhancement Funds - The CSI 300 index enhancement fund has achieved a year-to-date excess return of 8.1%, with a total return of 26.5% compared to the CSI 300's 18.4% [20] - The CSI 500 index enhancement fund has recorded a year-to-date excess return of 6.4%, with a total return of 33.2% against the CSI 500's 26.8% [24] - The CSI 1000 index enhancement fund has shown a year-to-date excess return of 13.8%, with a total return of 38.3% compared to the CSI 1000's 24.5% [31] Tracking Portfolio Performance - The report emphasizes the use of deep learning frameworks to construct the CSI 300, CSI 500, and CSI 1000 index enhancement portfolios, optimizing alpha signals and risk signals through a combination of multi-source features and neural networks [16][21][25][29] - The CSI 300 index enhancement portfolio has shown a total return of 26.5% year-to-date, outperforming the CSI 300 index by 8.1% [20] - The CSI 500 index enhancement portfolio has achieved a total return of 33.2% year-to-date, with an excess return of 6.4% [24] - The CSI 1000 index enhancement portfolio has recorded a total return of 38.3% year-to-date, with an excess return of 13.8% [31]