Using Tensorflow's LSTM (Long Short-term Memory) RNN (Recurrent Neural Network) to predict a cryptocurrencies next-day closing price
This notebook predicts the value of Bitcoin based on the 'Fear and Greed' index
This notebook predicts the value of Bitcoin based on a rolling 10-day window of closing prices
You can see, given the results, that the prediction based on closing prices has a smaller margin of error, averaging about 75% of the actual value in it's prediction, versus the Fear and Greed index at 61%.