Price Prediction of Ethereum Using Blockchain Historical and Exchange Data by Supervised Machine Learning Algorithms

Abstract

Predicting or forecasting crypto currency prices is now one of the most difficult tasks in crypto market trading due to its qualities and dynamic nature. The purpose of the present work is to analyse the exchange and blockchain data and develop a prediction model using machine learning. Ethereum (ETH) is one of the crypto currencies, and data has been taken from the price time series from January 1, 2017 to December 31, 2021, on a daily basis. The algorithm for gathering data has been trained and tested using a machine-learning algorithm. The adequacy of the developed machine learning models was validated using MAPE, RSME, MAE, and R2 scores. The developed model can predict future results with an accuracy of up to 85% for 7 days. Based on the findings, it is suggested that blockchain historical data and exchange data can be utilised as input characteristics in the development of a machine learning model to forecast Ethereum's future price.