This is the reinforcement learning code I used for my thesis about how to trade low market capitulization cryptocurrencies.
- Fetches up to date historical data from Binance, using a custom script.
- Plots a comparison of the reinforcement learning agent and simple trading strategies (see section Images for more info).
- Some extras, such as an analysis of all TA indicators available for the TA library.
The required packages to run this code can be found in the requirements.txt file. To run this file, execute the following code block:
$ pip install -r requirements.txt
Alternatively, you can install the required packages manually like this:
$ pip install <package>
- Clone the repository
- Run
$ python src/main.py - See result
After testing the RL agent a graph is plotted, showing the net worth of the agent compared to the benchmarks.
Displays a heatmap of absolute correlation of technical analysis indicators in the same group. This is how the heatmap of trend indicators looks like.
