Cryptocurrency forecasting with LSTM

Deep learning to forecast cryptocurrency prices 7 days in the future. Here are all the cryptos that have models: 'BTC','ETH','ADA','MATIC','DOGE','SOL','DOT','SHIB','TRX','FIL','LINK','APE','MANA',"AVAX","ZEC","ICP","FLOW","EGLD","XTZ","LTC"

Currently, there are 289 features (with pandas_ta, FRED data, and wikipedia data) to help with the forecasting of the close price

Usage

python3 crypto_deep_many_features.py all test #check the output for all cryptos
python3 crypto_deep_many_features.py all notest #perform training or create future prediction
python3 crypto_deep_many_features.py BTC test #check output for individiual crypto, in this case BTC.
python3 crypto_deep_many_features.py BTC notest #perform training or create future forecast for individual crypto, in this case BTC.
#change price over time
python3 change_price_over_time.py --name all --extension True #run all cryptos from a predefined list AND yfinance's trending list
python3 change_price_over_time.py --name BTC --extension True #run individual crypto
python3 change_price_over_time.py --name ^DJI --extension False #run individual stock

Forecasting percent change for all cryptocurrencies

Forecasting error

Forecasting examples of BTC

Relationship between Data Length and MAPE

Relationship between Sample Entropy and MAPE

Relationship between Fractal Dimension and MAPE

Month Analysis

Month

Rainbow Log plot of BTC

Shannon Entropy on the log returns on cryptos and stock market

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.