Prediction of Bitcoin Prices Based on Blockchain Information Using Long-Short Term Memory Method
The data used is all Bitcoin price data and Blockchain information. While the research sample is Bitcoin Price, Miners Revenue, Difficulty, Hash Rate, Median Confirmation Time, Average Block Size, Total Number of a Unique Bitcoint, Transaction per Block, Confirmed Transaction, and Cosh % per Transaction, each of these variables from period 04 August 2018 to 21 January 2020.
https://machinelearningmastery.com/multivariate-time-series-forecasting-lstms-keras/