🤖 AI-Cryptocurrency-Trader\n\nA revamped cryptocurrency trading bot leveraging Machine Learning and Rust. Originally inspired by a project from another developer.\n\n## ⚠️ Disclaimer\n\nThis bot is an experiment in enhancing an existing project using Rust and Machine Learning. The effectiveness of the trading strategy has not been fully explored. Utilize this program at your own risk!\n\n## 📖 Strategy\n\nThe trading strategy is relatively straightforward:\n\n- Gathers hourly kline (candle) data from Binance from the last X days.\n- Trains a machine learning model using the gathered data. The LightGBM framework is used, which is a swift gradient boosting framework that utilizes tree based learning algorithms for training. While the predictions aren't as accurate as other solutions like recurrent neural networks (RNN) like LSTM, my testing indicates it can be a good indicator for elementary market movements (price up or down), which is all that is required for this strategy.\n- Uses the trained model for predicting the current candle high price. If the predicted price is lower than the current open or close (i.e current) price, it waits for the next candle to start over. Otherwise, a buy order is placed.\n- Next, the bot waits until the prediction is realized. If the prediction isn't realized by the end of the candle, it will simply hold out until it is.\n\n## 💻 Installation & usage\n\nInstall Rust first, then clone this repository:\n\nbash\n$ git clone https://github.com/BuildCop/AI-Cryptocurrency-Trader.git\n$ cd AI-Cryptocurrency-Trader\n\n\nNext, copy the configuration file and adapt it (should be self-explanatory):\n\nbash\n$ cp config.example.yaml config.yaml\n$ vim config.yaml # or use any other text editor of choice to edit the config file\n\n\nTo run the bot in development mode, execute:\n\nbash\n$ RUST_LOG=debug cargo run\n\n\nTo run the bot in production mode, execute:\n\nbash\n$ RUST_LOG=info cargo run\n\n\nYou can also build a release binary with cargo build -r and copy it + your config file to a VPS or raspberry pi.\n\n## 📷 Screenshots\n