中国股票风险因子体系数据集

本文数据集摈弃了市值加权的回归分析方法，使用更为稳健的换手加权 EM 算法估测，并以此为基础，在申万行业因子之外臻选出 12 个具有强解释力且直观稳定的风格因子。我们使用 2012-2025 年中国 A 股上市公司股票的实际数据，使用模型进行实际风险控制校验，该因子体系相比于传统因子体系在控制波动率以及尾部风险均有显著提升，有助于制定更精准的风险控制与投资决策。

This dataset abandons the market-capitalization weighted regression analysis method, instead adopting a more robust turnover-weighted EM algorithm for estimation. Based on this framework, 12 style factors with strong explanatory power, intuitive and stable performance are carefully curated in addition to the Shenwan Industry Factors. We utilized actual data of stocks from public companies listed on China's A-share market covering the period from 2012 to 2025, and conducted actual risk control validation using the model. Compared with traditional factor systems, this factor system achieves significant improvements in both volatility control and tail risk management, and facilitates the formulation of more precise risk control and investment decisions.

搜集汇总

背景与挑战

背景概述

该数据集采用换手加权EM算法构建中国A股风险因子体系，基于2012-2025年全样本数据，在申万行业因子外精选了12个风格因子。实证表明，该体系在控制波动率和尾部风险方面优于传统方法，有助于提升风险控制与投资决策的精准性。