Predicting Global Computing Power of Blockchain Using Cryptocurrency Prices

Abstract

Blockchain is a disruptive technology that enables disparate users to share their information in blocks trustworthily without a centralized entity. One fundamental problem is how to stable the block interval. To address this problem, our method is: 1. predict the computing power (i.e., hashrate) of a blockchain system by the cryptocurrency price; 2. stable the interval according to the predicted power. This paper focuses on the prediction of the global computing power. In our prediction, we adopt a LSTM-based regression algorithm to handle the hysteresis of computing power changes in response to the price changes. Taking the Bitcoin system as an example, we run extensive experiments that verify that our prediction algorithm is very accurate.