Skip to content

README.md

Latest commit

 

History

History
266 lines (203 loc) · 8.73 KB

README.md

File metadata and controls

266 lines (203 loc) · 8.73 KB

QuantMarketLab

A comprehensive quantitative trading and market analysis platform featuring multi-factor models, technical indicators, pattern recognition, sentiment analysis, and backtesting capabilities.

🚀 Features

Core Functionality

  • Multi-Factor Model: Advanced machine learning models (XGBoost, LightGBM, CatBoost, Random Forest) for stock selection and prediction
  • Technical Indicators: Comprehensive technical analysis signals (RSI, MACD, CCI, Bollinger Bands, etc.)
  • Pattern Recognition: Automated pattern detection (Head & Shoulders, Double Tops/Bottoms, Triangles, etc.)
  • Sentiment Analysis: News sentiment analysis using FinBERT and RSS feeds
  • Backtesting Engine: Robust backtesting framework with transaction costs, slippage, and performance metrics
  • Factor Engineering: Custom factor generation and preprocessing (Winsorization, Z-score normalization, industry neutralization)
  • Data Management: Efficient data loading and caching with PostgreSQL integration

Modules

  • multifactor/: Multi-factor model implementation with modular architecture
  • backtester/: Backtesting engine for strategy evaluation
  • signals/: Technical signals and pattern recognition
  • factor/: Factor generation and engineering
  • util/: Data loading utilities for stocks, crypto, and news
  • news/: News sentiment analysis from RSS feeds
  • other_strategy/: Additional trading strategies and alerts

📋 Requirements

  • Python 3.8+
  • PostgreSQL (for data caching)
  • API Keys:
    • Tiingo API (for stock/crypto data)
    • Finnhub API (for fundamental data)
    • Hugging Face API (for FinBERT sentiment analysis)

🔧 Installation

  1. Clone the repository

    git clone https://github.com/Joe251/QuantMarketLab.git
cd QuantMarketLab

  2. Create a virtual environment

    python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

  3. Install dependencies

    pip install -r requirements.txt

  4. Install TA-Lib (if not already installed)

    # macOS
brew install ta-lib
pip install TA-Lib

# Linux
wget http://prdownloads.sourceforge.net/ta-lib/ta-lib-0.4.0-src.tar.gz
tar -xzf ta-lib-0.4.0-src.tar.gz
cd ta-lib/
./configure --prefix=/usr
make
sudo make install
pip install TA-Lib

  5. Set up PostgreSQL database

    • Install PostgreSQL
    • Create a database named postgres (or update db_params in code)
    • Update database connection parameters in app.py and other files

  6. Configure API keys

    • Update API tokens in relevant files:
      • Tiingo API token
      • Finnhub API key
      • Hugging Face token (for FinBERT)

📁 Project Structure

QuantMarketLab/
├── app.py                      # Flask API server
├── multi_factor.py            # Main multi-factor model wrapper
├── multifactor/               # Multi-factor model modules
│   ├── data_manager.py        # Data loading and management
│   ├── model_manager.py       # Model training and management
│   ├── preprocessor.py        # Factor preprocessing
│   ├── backtester.py          # Multi-factor backtesting
│   └── multifactor_model.py    # Main orchestrator
├── backtester/                # Backtesting engine
│   └── backtest.py            # Strategy backtesting functions
├── signals/                   # Technical signals
│   ├── tech_sig.py           # Technical indicators
│   ├── pattern_sig.py        # Pattern recognition
│   └── elliottwave_sig.py    # Elliott Wave analysis
├── factor/                    # Factor generation
│   ├── factor_engine.py      # Factor computation engine
│   └── factor_generator.py    # Factor generation functions
├── util/                      # Utilities
│   ├── stock/                # Stock data loading
│   ├── crypto/               # Crypto data loading
│   └── news/                 # News data and analysis
├── news/                      # RSS sentiment analysis
├── data/                      # Data storage
│   └── stock_metrics/        # Stock metrics data
├── models/                    # Trained models storage
└── logs/                      # Application logs

🚀 Quick Start

Running the Flask API Server

python app.py

The API server will start on http://localhost:5000 (or the configured port).

Example: Multi-Factor Model Training and Backtesting

from multi_factor import train_and_backtest_multifactormodel

# Train and backtest a multi-factor model
results = train_and_backtest_multifactormodel(
    load_fromcsv=True,
    train_model=True,
    train_startdate='2015-01-01',
    test_startdate='2019-01-01',
    train_enddate='2020-01-01',
    backtest_startdate='2020-01-02',
    backtest_enddate=None,
    rebalance_period=5,
    top_n=10
)

Example: Strategy Backtesting

from backtester import backtest_signal_strategy, plot_backtest_results
from signals.pattern_sig import PatternRecognizer
from util.stock.load_stock import load_stock_eodprice_from_tiingo

# Load data
df = load_stock_eodprice_from_tiingo('AAPL', '2021-01-01', '2023-12-31')

# Detect patterns
recognizer = PatternRecognizer('AAPL', df)
patterns_df = recognizer.analyze_all()

# Run backtest
df_backtest, summary, chart_data = backtest_signal_strategy(
    df=df.join(patterns_df),
    positive_signals_column=['positive_signal'],
    negative_signals_column=['negative_signal'],
    positive_signals_threshold=1,
    negative_signals_threshold=1,
    initial_capital=100000
)

# Plot results
plot_backtest_results(df_backtest, summary)

📡 API Endpoints

Backtesting

  • POST /api/backtest - Run backtest for a given ticker

Multi-Factor Model

  • POST /api/multifactor/run - Train and backtest multi-factor model

Sentiment Analysis

  • GET /api/sentiment/stock - US stock market sentiment
  • GET /api/sentiment/hkstock - Hong Kong stock market sentiment
  • GET /api/sentiment/ashare - A-share market sentiment
  • GET /api/sentiment/crypto - Crypto market sentiment

News Analysis

  • POST /api/news/analyze - Analyze company news sentiment

See app.py for complete API documentation.

🔬 Key Components

Multi-Factor Model

The multi-factor model uses machine learning to predict stock returns based on multiple factors:

  • Factor Preprocessing: Winsorization, cross-sectional Z-score normalization, industry/market cap neutralization
  • Models: XGBoost, LightGBM, CatBoost, Random Forest, Gradient Boosting
  • Backtesting: Rebalancing, turnover calculation, performance metrics (Sharpe ratio, drawdown, etc.)

Factor Engine

A flexible factor computation engine supporting:

  • Time-series functions (Ref, Delta, Mean, Std, etc.)
  • Cross-sectional functions (CSRank, CSZScore, etc.)
  • Technical indicators (RSI, MACD, CCI, etc.)
  • Custom factor expressions

Pattern Recognition

Automated detection of chart patterns:

  • Head & Shoulders
  • Double Tops/Bottoms
  • Triangles (Ascending, Descending, Symmetrical)
  • Flags and Pennants
  • And more...

📊 Data Sources

  • Stock Data: Tiingo API
  • Crypto Data: Tiingo API
  • Fundamental Data: Finnhub API
  • News Data: Tiingo News API, RSS feeds
  • Sentiment Analysis: FinBERT (Hugging Face)

⚙️ Configuration

Update configuration in relevant files:

  • Database: db_params in app.py and data loading modules
  • API Tokens: Update tokens in respective modules
  • Model Directory: Default models/
  • Data Directory: Default data/

📝 Notes

  • Ensure PostgreSQL is running and accessible
  • API keys are required for data fetching
  • Large datasets are cached in PostgreSQL for faster access
  • Model files are saved in models/ directory
  • Logs are written to logs/ directory

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

  1. Fork the repository
  2. Create your feature branch (git checkout -b feature/AmazingFeature)
  3. Commit your changes (git commit -m 'Add some AmazingFeature')
  4. Push to the branch (git push origin feature/AmazingFeature)
  5. Open a Pull Request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

⚠️ Disclaimer

This software is for educational and research purposes only. It is not intended as financial advice. Trading involves risk, and past performance does not guarantee future results. Always do your own research and consult with a qualified financial advisor before making investment decisions.

📧 Contact

For questions or support, please open an issue on GitHub.

Made with ❤️ for quantitative trading enthusiasts