Combining Blockchain and Machine Learning to Forecast Cryptocurrency Prices

Abstract

Blockchain is an emerging technology that enables a vital framework for various cryptocurrency operations such as bitcoin. Notably, without any involvement from third party authorities, blockchain offers a decentralized consensus scheme to process user transactions, fund transfer, and various data records in a secure and reliable way. Furthermore, bitcoin price forecasting has been a vital research trend, where machine learning techniques play a substantial role. A sophisticated and appropriately trained model can be useless if the features being tested are unreliable. Independently, one of the most desirable aspects of a system that utilizes the blockchain is the concrete objectiveness by which each entry is cataloged. Any data collected and reported on the blockchain is unambiguous, and therefore, extremely suitable for a machine learning algorithm. To efficiently forecast bitcoin price movements, in this work, we propose and examine various lenses by which to view this union, each with varying degrees of success.