Bitcoin-Trader-ML

Bitcoin Trader ML is a automatic crypto trading bot for Coinbase using Transformer Neural Networks.

Train one Autoformer for crypto price predictions, and one for trade decisions (-1: Sell, 0: Hold, 1: Buy).

Requirements Installation

To get started with this project, you need to install the required packages. Follow the steps below:

1. Create a virtual environment (optional but recommended): bash python -m venv venv source venv/bin/activate # On Windows, use `venv\ScriptsActivate`

2. Install the required packages: bash pip install -r requirements.txt

Usage Instructions

1. Clone the repository: bash git clone https://github.com/bradym05/Bitcoin-Trader-ML.git cd Bitcoin-Trader-ML

2. Get your coinbase API key:

   Create your API key and download the JSON file (coinbasecloudapi_key.json) https://portal.cdp.coinbase.com/access/api

3. Setup private folder:

   In Bitcoin-Trader-ML\main Create folder "private" Copy your coinbase API file to Bitcoin-Trader-ML\main\private\coinbasecloudapikey.json Create init file inside of "private" folder: ```python keyfilepath = "main\private\coinbasecloudapikey.json" portfolio_uuid = "YOUR PORTFOLIO UUID" ```

4. Train Price Prediction Autoformer:

   Using neuralforecast,

   1. Preprocess your dataset:

      Columns

      y is the price python ['ds', 'unique_id', 'open_price', 'high', 'low', 'y', 'btc_vol', 'usdc_vol', 'ema0', 'ema1', 'ema2', 'ema3', 'ema4', 'ema5', 'ema6', 'ema7', 'ema8', 'ema9', 'wma0', 'wma1', 'wma2', 'wma3', 'wma4', 'wma5', 'wma6', 'wma7', 'wma8', 'wma9', 'sma0', 'sma1', 'sma2', 'sma3', 'sma4', 'sma5', 'sma6', 'sma7', 'sma8', 'sma9'] Example Dataframe bash ds unique_id open_price high low y btc_vol usdc_vol ... sma2 sma3 sma4 sma5 sma6 sma7 sma8 sma9 0 2017-08-19 00:00:00 0 0.143153 0.145590 0.142314 0.145074 0.001439 5.925526 ... 0.149302 0.148726 0.148141 0.146973 0.145702 0.144943 0.144299 0.143965 1 2017-08-19 00:05:00 0 0.143313 0.145609 0.142485 0.145030 0.001403 5.780747 ... 0.149254 0.148677 0.148043 0.146867 0.145639 0.144889 0.144271 0.143939 2 2017-08-19 00:10:00 0 0.143473 0.145627 0.142655 0.144987 0.001367 5.635968 ... 0.149206 0.148628 0.147946 0.146761 0.145576 0.144836 0.144243 0.143914 3 2017-08-19 00:15:00 0 0.143633 0.145646 0.142826 0.144943 0.001331 5.491189 ... 0.149158 0.148579 0.147849 0.146655 0.145512 0.144782 0.144216 0.143888 4 2017-08-19 00:20:00 0 0.143793 0.145664 0.142997 0.144900 0.001294 5.346409 ... 0.149110 0.148531 0.147751 0.146549 0.145449 0.144728 0.144188 0.143862 [648853 rows x 38 columns]

   2. Train your autoformer:

      Optional: Edit and use machinelearning\training\NFAutoformer.py according to your dataset

      Save your trained autoformer to: machinelearning\training\saved\Autoformer30M\

5. Train Decision Autoformer:

   Using neuralforecast,

   1. Preprocess your dataset:

      Decisions are mapped as follows:

      Decisions Chart

      | Value | Decision | |----------|----------| | -1 | Sell | | 0 | Hold | | 1 | Buy |

      Columns

      y is the decision value p1-p12 are predictions from the price prediction autoformer python ['y', 'ds', 'unique_id', 'open_price', 'high', 'low', 'close_price', 'btc_vol', 'usdc_vol', 'ema0', 'ema1', 'ema2', 'ema3', 'ema4', 'ema5', 'ema6', 'ema7', 'ema8', 'ema9', 'wma0', 'wma1', 'wma2', 'wma3', 'wma4', 'wma5', 'wma6', 'wma7', 'wma8', 'wma9', 'sma0', 'sma1', 'sma2', 'sma3', 'sma4', 'sma5', 'sma6', 'sma7', 'sma8', 'sma9', 'p1', 'p2', 'p3', 'p4', 'p5', 'p6', 'p7', 'p8', 'p9', 'p10', 'p11', 'p12'] Example Dataframe bash y ds open_price high low close_price btc_vol usdc_vol ... p4 p5 p6 p7 p8 p9 p10 p11 p12 0 1 2017-08-19 00:00:00 0.143153 0.145590 0.142314 0.145074 0.001439 5.925526 ... 0.144832 0.14463 0.144239 0.143678 0.143814 0.14406 0.144468 0.144901 0.142082 1 1 2017-08-19 00:15:00 0.143633 0.145646 0.142826 0.144943 0.001331 5.491189 ... 0.144832 0.14463 0.144239 0.143678 0.143814 0.14406 0.144468 0.144901 0.142082 2 1 2017-08-19 00:30:00 0.144113 0.145701 0.143339 0.144813 0.001222 5.056851 ... 0.144832 0.14463 0.144239 0.143678 0.143814 0.14406 0.144468 0.144901 0.142082 3 1 2017-08-19 00:45:00 0.144594 0.145757 0.143851 0.144683 0.001113 4.622513 ... 0.144832 0.14463 0.144239 0.143678 0.143814 0.14406 0.144468 0.144901 0.142082 4 1 2017-08-19 01:00:00 0.145074 0.145812 0.144364 0.144552 0.001005 4.188176 ... 0.144832 0.14463 0.144239 0.143678 0.143814 0.14406 0.144468 0.144901 0.142082 [216285 rows x 51 columns]

   2. Train your autoformer:

      Optional: Edit and use machinelearning\training\NFDecisionAutoformer.py according to your dataset

      Save your trained autoformer to: machine_learning\training\saved\DecisionAutoformer\

      Note: You must convert your autoformer's raw outputs into int decision values

6. Begin Auto Trading:

   Example Usage: ```python from main import AutoTrader

   Create AutoTrader object, set paper to true to simulate trading

   auto_trader = AutoTrader(paper=True)

   Wait for auto trader to make decisions

   time.sleep(5)

   Print decision history

   print(auto_trader.predict.decisions) ```

Authors

| Author | Contact Info | |-------------|----------------------| | Brady Maki | |

📜 License

Bitcoin Trader ML is licensed under the MIT License. See the LICENSE file for more information.