Market-Pattern-Analysis-using-machine-learning

Market Pattern Analysis using machine learning

This project is a Python-based quantitative research workflow for testing whether a machine-learning model can predict short-term market direction using technical, candlestick, macro, and sentiment-style market features.

The main script, pattern_ml_backtest_V4_sentiment.py, downloads historical market data, engineers features, trains an XGBoost classifier, and runs a walk-forward backtest to evaluate predictive performance.

What the project does

• Downloads OHLCV data from Yahoo Finance
• Builds technical indicators such as RSI, MACD, ATR, moving averages, volatility, and momentum
• Adds candlestick pattern features such as doji, hammer, shooting star, and engulfing patterns
• Incorporates macro-market features including VIX, DXY, SPY, Fed funds, yield curve, and CPI inflation data
• Trains a binary classifier to predict whether the next bar will be up or down
• Runs a walk-forward backtest and reports metrics such as win rate, drawdown, Sharpe ratio, and total return
• Saves a trained model artifact for later use

Project file

• pattern_ml_backtest_V4_sentiment.py — main training, feature engineering, backtest, and model export workflow

Requirements

Install the required Python packages:

pip install yfinance pandas numpy scikit-learn xgboost fredapi joblib --break-system-packages

Usage

Run the sentiment-enhanced backtest script:

python pattern_ml_backtest_V4_sentiment.py

This script will:

1. Load market data for the configured tickers
2. Build features from price and macro data
3. Train the model
4. Print backtest metrics and save the trained model file

Notes

• The script is intended for research and experimentation, not live trading
• The workflow uses Yahoo Finance and FRED data, so internet access is required
• Model performance depends heavily on market regime, data quality, and feature tuning
• The script currently targets example tickers such as GC=F and EURUSD=X

Disclaimer

This repository is for educational and research purposes only. It does not provide financial advice and should not be used as a standalone trading system without further validation.