Awesome-AI-Trader

This is a curated automated AI app for quantitative trading for stocks, futures and options. intergrated data fetching before it, and backtesting after it. For data, it taks advantage of Yahoo finance (For daily price data) and CFTC (for Commitment of Traders data).

• Awesome-AI-Trader
  • Project proposal / design
  • Pipeline
  • To do list
  • Local run: Quick Start
    • Requirement
    • Start a demo
  • License

Project proposal / design

This project is trying to automate the data manipulation, AI tuning for quantitative trading, and provide good user experience like visualization. Its objective is to give a good buy/sell signal to user.

For details, please find the Project proposal / design in documentation.

Pipeline

1. Data fetching: Fetch data from Yahoo finance, CFTC, and other sources.
2. Data manipulation: Clean and manipulate data: fill missing data, normalize data, etc.
3. Feature extraction: Extract features from data, like moving average, RSI, etc.
4. Preprocessing: Use MIC, Ramdom Forest, etc to get the best features for AI model.
5. Train AI model: Use LSTM, CNN, etc to predict the price.
6. Backtesting: Backtest the AI model with historical data.
7. Visualization: Visualize the backtesting result.
8. User interface: Provide a good user interface for user to interact with the app.

To do list

• Adpot constant strategy when training AI.
• Compute whole market infomation and add this feed into every single stock data.

Local run: Quick Start

Though This project is oriented to serve as some beckend for a potential application in the future, user can still run and test it locally.

Requirement

User is expected to have docker installed in their computer.

Start a demo

After you download this project, open a terminal (git.bash if you are using Windows). Start docker service, if not, and run the following:

./deploy.sh

And this app will be built, start ,and work in the terminal.

License

MIT