Onour AI Trading Bot
This project is a sophisticated, autonomous cryptocurrency trading bot designed to operate on the Binance exchange. It employs a cutting-edge Deep Reinforcement Learning strategy to make intelligent trading decisions. The core of the bot is a Duelling Deep Q-Network (DQN) enhanced with a Long Short-Term Memory (LSTM) layer, allowing it to effectively learn from sequential market data and adapt to changing conditions. The bot connects to Binance using WebSockets for real-time data, ensuring it always has the most current information. Key features include fully autonomous trading, dynamic trade sizing based on the model's confidence, detailed logging of all orders for performance review, and state management that allows the bot to be stopped and restarted without losing its learned progress. Built with Python, it is designed for both testing on the Binance Testnet and deployment in live trading environments.
Category tags:"Ahmed, Awesome presentation. Feels a bit overstated though. I'm concerned about legal, and how the bot actually works. I pulled the repo installed libs, ran python3 tui.py, got the exchange pricing table. Would also love to see the app working in some bits (beyond the screens in the repo/this 1 command), and also could not find the 'sophisticated' RL in the repo. Indeed, The bot.py contains a DQN (Deep Q-Network) implementation, but I would classify it as basic to intermediate rather than advanced. There's simple LSTM and indicators. You may want to continue working on the implementation and compliance/UX to cater to actual users, and in finance (even crypto) please avoid overstating/overpromising, you want to build trust. Other than that, fingers crossed!"
Mark Bain
For Innovators & Creators