charliedream1/ai_quant_trade

🤖 AI Quantitative Trading Bot – Quick Overview

What it is – A large, community‑driven repository that bundles together everything you need to learn, prototype, and (optionally) run quantitative‑finance strategies that use AI techniques. It covers the whole pipeline:

• Education – markdown notebooks and notes on finance theory, data sources, and AI basics.
• Data handling – scripts for pulling market data from Wind, akshare, Eastmoney, etc., with a unified interface.
• Alpha mining – libraries of factor formulas (alpha101, ta‑lib, stockstats) and a demo that auto‑mines 5 000 alphas using tsfresh.
• Strategy examples – both classic rule‑based (double‑moving‑average, portfolio optimisation) and AI‑heavy approaches (reinforcement‑learning agents, GNN, deep‑learning, LLM‑driven forecasts).
• LLM applications – notebooks and training scripts for a “reasoning stock‑forecasting” model built on the Unsloth framework, plus a tutorial on using LLMs for market analysis.
• Live‑trading helpers – Excel‑based monitoring tools, a Streamlit market dashboard, and integration with the Wind platform for institutional‑grade simulation.
• Resource hub – a curated, reviewed collection of thousands of external papers, codebases, and tutorials, organised by the full quant workflow.

Who it’s for – Institutional quants, retail programmers, and even non‑programmers who want a guided path (the repo includes low‑code Excel tools).

How to get started – Clone the repo, install the requirements.txt, then dive into the egs_* folders that each contain a README with step‑by‑step instructions. For example, run the RL tutorial with:

cd egs_trade/rl/a002_finRL_tutorial/a01_Stock_NeurIPS2018
python run.py

Key take‑aways

• One‑stop platform covering data, research, backtesting, and deployment.
• Supports multiple markets (stocks, funds, crypto) and deployment targets (Python, C++, CPU/GPU).
• Actively maintained with recent features (2026‑07‑25 “Stealth Stock Trading” Excel tool, 2025 LLM‑based forecasting).
• Not a packaged library – you explore examples rather than import a single pip package.

---

📂 Repository Structure (high‑level)

ai_quant_trade/
├─ ai_notes/            # Theory, papers, tutorials
├─ docs/                # Repo‑wide docs
├─ egs_aide/            # Monitoring & assistant tools
├─ egs_alpha/           # Alpha mining & factor libraries
├─ egs_data/            # Data acquisition wrappers
├─ egs_fin_nlp/         # NLP‑based sentiment tools
├─ egs_llm/             # LLM training & inference examples
├─ egs_online_platform/ # JoinQuant / Uqer scripts
├─ egs_trade/           # Strategy demos (RL, vanilla, Qlib, etc.)
├─ quant_brain/         # Core algorithm utilities
├─ runtime/             # Deployment helpers
├─ tools/               # Misc utilities
└─ requirements.txt

---

🔧 Typical Use‑Case Flow

1. Learn – Read ai_notes markdown/Jupyter notebooks.
2. Gather data – Use scripts in egs_data to pull price, news, sentiment.
3. Mine alphas – Run the tsfresh auto‑mining demo in egs_alpha/auto_alpha/tsfresh.
4. Build a strategy – Start from a vanilla double‑MA example or the RL tutorial.
5. Backtest – Each example ships with its own backtest runner and performance metrics.
6. Deploy – Follow runtime instructions to run live (Wind simulation, Python, or C++).
7. Assist – Keep an eye on markets with the Excel “Stealth Stock Trading” tool or the Streamlit dashboard.

---

📚 Resources & Community

• A massive curated list (a_全网优秀资源) of external open‑source projects, papers, and tutorials.
• Video tutorials and a WeChat “Knowledge Planet” for hands‑on walkthroughs.
• The repo is Apache‑2.0 licensed, so you can reuse code in commercial or research settings.

---

Bottom line – This repo is a genuine, extensive open‑source effort that brings together data pipelines, AI‑driven strategy prototypes, and practical trading utilities for quantitative finance.