Download
SGX-Full-OrderBook-Tick-Data-Trading-Strategy is an open-source research project focused on modeling high-frequency financial market behavior using machine learning techniques. The repository analyzes tick-level order book data from the Singapore Exchange and attempts to capture the dynamics of limit order book movements. By extracting features such as order depth ratios and price movement indicators, the system trains machine learning models to predict short-term market changes. Several algorithms are used during model selection, including Random Forest, Extra Trees, AdaBoost, Gradient Boosting, and Support Vector Machines. The project evaluates models by predicting price direction within very short time windows and then applying a simple trading strategy based on those predictions. It also measures profitability through profit-and-loss analysis derived from the predicted signals.
Features
- Framework for modeling high-frequency limit order book dynamics
- Feature engineering techniques such as rise ratio and depth ratio indicators
- Multiple machine learning algorithms for model training and comparison
- Prediction of short-term price movements using tick-level data
- Backtesting pipeline for evaluating trading strategies and profit outcomes
- Visualization tools for analyzing prediction accuracy and trading performance
Project Samples
