Skip to content

[收录申请] skill-factor-mining-pandaai — 论文因子挖掘与回测分析 #1

Description

@TerribleCookie

申请收录 Skill:factor-mining-pandaai

基本信息

项目 内容
仓库 TerribleCookie/skill-factor-mining-pandaai
名称 factor-mining-pandaai
类型 skill
申请分类 02 因子研发工具箱

功能描述

从本地研报/论文 PDF 中自动提取量化因子表达式,并通过 pandaai-cli 完成创建、回测、分析的全流程。

适用场景:

  • 用户上传研报/论文/文档,要求从中挖掘量化因子
  • 将文档中的数学公式/表达式转换为可执行的量化因子
  • 对提取的因子进行回测/分析/提交参赛

核心指标(测试论文实例)

以 Cross Impact in Derivative Markets(SSRN-3779461)为测试论文,默认回测窗口回测结果:

指标 数值
IC_mean 0.0316
多空组合年化 44.04%
多空组合夏普 1.57
最大回撤 11.50%
月度胜率 100%
单调性 0.67

文件结构

skill-factor-mining-pandaai/
├── SKILL.md                    # 核心定义(含 quantSkills frontmatter)
├── README.md
├── factor-mining-pandaai.skill # 打包后的 skill 文件
├── references/
│   └── pandaai_cli_reference.md
└── scripts/
    └── pandaai_cli_wrapper.py

技术栈

  • Python >= 3.10
  • pandaai-cli(PandaAI 因子分析平台)
  • pdfplumber(PDF 文本提取)

数据来源

PandaAI 平台提供的 A 股日频行情数据(open/high/low/close/volume)

已知限制

  1. 平台公式语法不支持行业中性化、截面标准化(需线下处理)
  2. Python 代码因子(--code)存在兼容性问题,建议优先使用 --formula
  3. Windows 环境需加 --json 参数避免编码错误
  4. 默认回测期较短(约 2 个月),建议手动指定 --start-date 延长

风险声明

本项目仅供研究和教育目的,不构成投资建议。因子回测表现取决于论文内容与 A 股市场的匹配度,不同论文效果差异较大。

元数据

quantSkills:
  category: factor
  tags: [factor-mining, pandaai, paper-extraction, backtest, quantitative]
  platforms: [claude-code, codex, openclaw, cursor, workbuddy]
  status: active
  validation_level: L1
  maintainer_type: community
  summary_zh: 从研报/论文中自动提取量化因子,并通过 PandaAI 完成回测分析
  summary_en: Auto-extract quantitative factors from research papers and backtest via PandaAI

请求

请审核并收录到 02 因子研发工具箱 分类下。如需修改或补充材料,请告知。


Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions